Answers.

169 verified answers to the questions planning, partner, IT, security and procurement teams ask about Kelynto. Each one was checked against the product's code or documents, and each carries a label: available today, available through enterprise implementation, planned, or not available.

Questions about Kelynto? Ask us.

Every answer is listed below, and each has a page of its own. For anything else, email marc@kelynto.com or book a 20-minute conversation.

Product

How is Kelynto different from a BI dashboard? Available today

A dashboard shows what happened and where: sales against plan by store and category, as a chart or table. It leaves the why to the reader. Kelynto is not a BI tool and does not try to be one: there is no free-form analysis, no custom chart builder. What it adds is the step after the dashboard: it combines your variance with in-stock, promotions and local signals near each store, shows which recorded causes carry an estimated impact and how much of the shortfall is not explained, drafts the written review, and records whether your planner agreed. Keep your dashboards; they answer different questions. The labeled findings can be exported as CSV and loaded into your BI tool.

Could our BI or data science team build this ourselves? Available today

Your team could, and what the work involves is specific: matching weather, alerts, events and road work to each store by distance and local week; separating a store-wide traffic effect from a category effect; judging a promotion or event against normal variation so noise is not credited; drafting text whose numbers are checked; and capturing planner verdicts as labels. A fair comparison is time: run a few of your past weeks through Kelynto and set the drafts beside what your team wrote.

Is Kelynto a chatbot or a wrapper around a language model? Available today

No to both. Kelynto is variance commentary for retail planning, not a chat product, and the engine that explains the week calls no language model at all. The analysis (the variance breakdown, cause attribution, the Watch list) is a deterministic rules engine working on your weekly plan, actuals, in-stock, promotions and local signals, and the built-in commentary writer needs no model. A planner edits, decides and publishes. The AI writer is switched on per workspace, by its admin or under a partner's recorded agreement with its clients; it writes from the week's evidence, every figure is printed from that evidence, and a claim that fails its check is dropped. Inside the product an assistant answers a fixed set of questions about your own workspace's records by rule, naming the record behind each sentence. The assistant on this website is separate and has no access to any workspace.

How is Kelynto different from doing this in Excel? Available today

Excel is fine for the numbers, and Kelynto accepts Excel and CSV files as its input. The difference is the work after the numbers. In a spreadsheet, a planner still has to find out whether a miss was a stockout, a storm, road work, a local event or a competitor opening, usually by checking several other sources, and then write it up. Kelynto does that lookup for every store and category each week, drafts the explanation with the evidence attached, and keeps a record of which causes the planner confirmed. It also keeps versions and an audit trail, which a shared workbook does not.

Is Kelynto a forecasting tool? Available today

No. Kelynto does not replace your planning system's forecast or plan, and it does not replenish, allocate, price or order. Kelynto is variance commentary for retail planning: it takes the plan you already have, compares last week's actuals with it by store and category, and explains the difference with the evidence behind every cause. Looking ahead, the Outlook sets what is coming beside your own plan or latest forecast, by cause: promotions, store closures, holidays and stock-out risks, sized from your own history and conditional on what is on file. A planner marks each item, and the export carries those decisions to your forecasting system; Kelynto changes nothing in it. Kelynto makes no claim to improve forecast accuracy: the Outlook's look-back on your own weeks is the measure. If your team explains variance to plan by hand every week, this is the part it covers.

How is Kelynto different from asking a general AI chatbot? Available today

A general AI chatbot will write a fluent explanation of any numbers you paste in, but it has no evidence about your stores and nothing checks what it says. Kelynto works differently in four ways. The causes come from a rules engine working on your weekly data and local signals matched to each store, not from a language model. Every number in the draft is checked against your data, and a draft with an unmatched number is discarded. What the records do not explain is stated as an amount. And the planner's verdict on each cause is recorded and scored, so you can see whether it is earning trust. The assistant inside the product follows the same rule: it answers from your workspace's records by fixed rules, and each sentence names the record it came from.

How is Kelynto different from Blue Yonder, RELEX, o9, SAP IBP, Oracle or Anaplan? Available today

The same answer applies to each of them, because the difference is one of category, not features. Blue Yonder, RELEX, o9, SAP Integrated Business Planning, Oracle's retail planning products and Anaplan are planning and forecasting platforms, by their own public descriptions. Kelynto produces neither. It takes the plan you already have, explains last week's variance to that plan by store and category with evidence, and learns from what your planner confirms or corrects. We do not compare features with any of them and make no claim about what they can or cannot do; your own team knows your implementation best. You keep your planning system. Kelynto works from weekly files, which any planning system able to export them can supply, and there is a file recipe for each of these systems. If your team still writes a Monday variance explanation by hand, the two sit side by side.

How is Kelynto different from a planning system? Available today

They do different jobs. A planning system produces the plan and the forecast, and often the orders that follow. Kelynto starts where that ends: it takes the plan you already have, compares last week's actual sales with it by store and category, and sets the variance beside the evidence: causes your own records show, with an estimated impact, and what a planning system does not hold, such as a storm, road work, a local event or a competitor opening near a store, listed as context. Then it records what your planner confirmed or corrected. It does not produce your plan or replace your forecast, and it writes nothing back; its Outlook sets what is coming beside your own plan or latest forecast, by cause, for your planner to decide on. You keep your planning system. Kelynto works from weekly files exported from it, so it sits above whichever system, or spreadsheet, produces your plan.

Does Kelynto replace our planning system? Available today

No. Kelynto replaces nothing. It does not produce your forecast or plan, replenish or order, and it does not write back to any system. Your planning system keeps producing the plan; Kelynto reads a weekly export of sales against that plan and explains the variance. Nothing is migrated and nothing in your current process is switched off, so if it does not help, you stop and nothing has changed. The only thing it aims to take over is the manual work of finding out why a store or category missed plan and writing that up for leadership, and even there a planner still edits and publishes every review.

Why isn't my current planning system already doing this? Available today

It may be, in part, and you should check. Planning suites are built to produce the plan and forecast, and many include exception reports, alerts or analytics on forecast error. We make no claim about what any vendor's product can or cannot do. The practical test is simple: does a planner on your team still spend time each Monday working out why stores missed plan, looking at weather, local events, road work or store issues, and writing a paragraph for leadership? If yes, that work is what Kelynto covers: a drafted explanation with evidence, a planner's verdict on each cause, and a count, from real use, of how useful that was. You keep your planning system either way.

Can a store manager see only their own district or region? Available today

Yes. An admin can limit a member, an API key or a webhook to the stores of some regions, districts or stores, and the limit can follow your identity provider's groups. A limited member reads those stores only: every page, document, email, share link, export, API answer and assistant answer is about them, with none of the company's totals, and they do not label, edit or publish the review. A limited planner can record a cause and an action for their own stores, seen by your planners. A colleague who emails them a review sends them the review of their own stores, while everyone else on the message gets the review as chosen. The limit is enforced in the database and in the application, tested on every surface. Admins always read the whole workspace.

How accurate is Kelynto? Available today

There is no accuracy figure on retailer data yet, so none is quoted. What holds on every review: every figure reconciles to the week's variance to plan; a cause gets an amount only when your own records show it; every amount is labelled an estimate and opens to the record behind it; weather, local events, road work and competitor openings are shown on file with no amount; and what the records do not explain is stated. Automated tests hold that arithmetic on generated weeks. The demonstration workspace is synthetic, and its numbers are not results. The measure that counts is on your own weeks: a pilot starts with eight sealed weeks of your history, scored against what two of your planners judged separately and beside simple baselines, then planner agreement and the edit rate on live weeks. Planner agreement is agreement between a planner and the review.

Can a planner add a cause the review missed? Available today

Yes. A planner can record a cause the review did not have, at any level: its kind, what happened, the stores and dates, and optionally a link, a file and their own estimate. It is a human statement, with who added it and when, under "Added by planners" on screen and in a section of its own in every document and email. An estimate is labelled "Planner's estimate" and never enters the arithmetic: the amounts are the same before and after.

A store or district manager who reads only their own stores can leave a short dated note for one of them (store voice): a reset, a staffing gap, a local event, from a phone. It shows as context with who wrote it and when, never with an amount; a planner can record a cause from it.

Which causes can Kelynto identify? Available today

The review has three parts. First, causes recorded in your own files carry an estimated impact: out of stocks, store closures and promotions that ran. Second, weather, local events, road work and competitor openings near your stores are listed as context on file, with place and dates and no amount. Kelynto does not credit them with a share of the shortfall today; a factor moves from context to a credited cause only when its effect has been shown on independent data. Third, every shortfall no recorded cause is credited with is reported as not explained, every week. Weather, alerts, events and road work can be pulled from outside sources; competitor and store-operations records come from a file you send. A planner can say a record on file mattered, or name a pricing change, a plan error or a data issue. Kelynto does not model cannibalisation, halo, price elasticity or assortment changes.

Is the commentary written by a template or by AI? Available today

Either. The engine computes every number, and the words are written by Kelynto's built-in writer or by the AI writer. The built-in writer needs no outside AI service and is always there as the fallback. The AI writer is switched on for a workspace by its admin, or for a partner's clients under the partner's recorded agreement with them: Anthropic Claude by default, or the company's own model (Anthropic, OpenAI, Microsoft Azure OpenAI, xAI Grok, Google Gemini or any OpenAI-compatible endpoint) with its own key. The model reads the week's evidence and cites it; it never types a number, and Kelynto prints each figure. Every claim is checked against what it cites and dropped if it fails; if the text fails as a whole, the built-in draft is kept with the reason. A planner reads, edits and publishes the review.

Can we compare one week with another? Available today

Yes. The Compare page puts any two weeks side by side at any level: the whole workspace, a region, a store or a category. It shows the figures of each week, the recorded causes with their estimates, the records on file and what planners decided, and marks what differs. A second view lists the last weeks of one level, from 4 to 26, each beside the week before it.

Each week is shown exactly as its own review states it. The page marks a difference as a difference and does not say why a number moved. A week with no sales, no review, or nothing at the chosen level says so, and the other week is still shown.

How does Kelynto handle currency? Available today

Currency is a label per workspace, not a conversion. Amounts are loaded exactly as you send them and printed with the mark of your currency setting: the dollar sign for USD and CAD, the pound and euro signs for GBP and EUR, and any other three-letter code as a prefix. Numbers are written in English format. If the optional AI writer produces an amount with a different currency symbol, the draft is rejected. Two limits: there is one currency per workspace, so a chain reporting in two currencies would need to send one of them converted, and changing the setting later does not restate reviews already written. Currency symbols and thousands separators in your files are read correctly on upload.

Is the demo data real? Available today

No. Harborview Markets is a made-up retailer: 36 stores and 26 weeks of synthetic sales, plan, in-stock and signals produced by a generator. The shocks in it were written by that generator, so the engine finding them shows that the pipeline works, not that it is accurate. Every review in it is a draft for you to edit and label, and its Score and learn page counts only real use, so it shows a made-up illustration, labelled as one, of how the page reads after eight weeks of use. You can confirm, reject and correct causes in it to see how review works. Screenshots on the site come from this workspace. None of its numbers are results and none should be quoted as such. No real retailer's data, name or logo appears anywhere in the product or the site.

Can the review documents carry our logo and layout? Available today

Yes. Each workspace has its own template, set by a workspace admin. It chooses which sections a document holds and in what order, an accent colour, a header line and a footer line, the default audience, and a logo. The summary, which states the amount that is not explained, cannot be removed. The screen shows a preview of your latest review as you change the settings. Every save is kept as a version and an earlier one can be restored. A logo must be a PNG or JPEG image of at most 1 MB; other formats, SVG among them, are refused, and Kelynto re-encodes the image and keeps only its own copy. A template does not change fonts or page layout, and documents are in English on US Letter pages.

Can planners edit the draft, and are versions kept? Available today

Yes. The draft is the first version; every planner save creates a new version, and the app keeps the history, can show a planner's text against the original draft, and can restore an earlier version. If two people save at once, the second gets a clear conflict message instead of silently overwriting. An edit in progress is held in the browser tab, so it survives a sign-in that expires mid-edit and is offered back afterwards. The difference between the draft and the planner's final text is measured word by word as the edit rate: 0 means the draft was sent unchanged, 1 means it was rewritten. That number is one of the three measures with a target on the Score and learn page. Every decision on a review (labels, added causes, edits, publishing) is also kept with who made it and when, and that history can be downloaded.

How does Kelynto decide what caused a variance? Available today

It starts from your own records, not from a guess. The review credits an estimated amount only to a cause your files record: a stockout, read from the drop in in-stock against the store's own recent level; a store closure; or a promotion that ran, stated with the stores and category your calendar gives it. Each estimate is shown with the records behind it and is labelled an estimate. Weather, local events, road work and competitor openings are matched to stores by distance and week and listed as context on file, with no amount. Whatever the recorded causes do not cover is stated as an amount, and every figure reconciles to the week's variance to plan. The engine is rules, not a trained model and not a language model. A planner confirms or corrects every item, and those verdicts are counted on Score and learn.

How does Kelynto show how well a cause is supported? Available today

By showing the record behind it, one click away. A recorded cause, such as a stockout, opens to the records it rests on: which record, which dates and which stores, for example the in-stock file with each store's rate that week against its usual rate. Its amount is labelled an estimate, because a record confirms that something happened and not the size of its effect. A weather alert, a local event, road work or a competitor opening is listed as context on file with its place, its dates and the record it came from, and no amount. What no recorded cause covers is reported as not explained, and that amount is always on screen. There is no confidence percentage. The planner reads the evidence and confirms, rejects or corrects.

What is the Explain, Watch, Act, Score loop? Available today

Four steps, repeated every week.

Explain: last week's sales against plan by region, store and category, in three parts: causes your own records show, each with an estimated impact and the records behind it; weather, events, road work and competitor openings on file near your stores, with no amount; and the share of the shortfall that is not explained.

Watch: one page lists what is on file near your stores for the coming weeks (road work, events, weather, competitor openings), each with its store, its dates and where the record came from. No sales figures are forecast.

Act: the planner edits the draft, confirms, rejects or corrects each recorded cause, says whether a record on file mattered, marks Watch items and publishes the review.

Score: the Score and learn page, built only from real use: draft edit rate, planner agreement (never called accuracy) and whether planners would use it every week.

What can we export from Kelynto? Available today

The weekly review downloads as a Word document, a PDF or a PowerPoint deck, at the level that is open: the whole company, a region, a store or a category, as a leadership summary or in full detail. Each states the amount that is not explained. For someone who uses a screen reader, send the Word file; the PDF is untagged. From a published review, the recap leadership already receives downloads as a workbook (the week in three parts per region, store and category, the week by store and category, decisions and follow-ups), as slides and as email text; every figure in them is the review's. Analysts and admins can download a reporting export: seven flat tables as CSV or Parquet, with a documented schema and no score. You own this data.

Can we track follow-up actions from the review? Available today

Yes. From a review you can add a follow-up action about a recorded cause, a record on file, a movement that is not explained, or the review as a whole. Each action has a title, an owner who is a member of your workspace, a due date and a status: open, in progress, done or dropped. It keeps a link back to the place in the review it came from. The Actions page lists them with filters for your own, overdue, a store, a week and a status, and remembers your filter. Every change is kept in a history. Planners, analysts and admins add and edit actions; an owner can move the status of their own. Actions appear in the review and in every document made from it. Owners get a reminder for actions due or overdue, by email where a mail relay is configured.

What are the limits of Watch? Available today

Three things to know about Watch. First, it lists and does not predict: an item is a record on file near a store, and it states no amount and no direction for sales, by design. Second, the horizon is short: a review's list looks three weeks ahead by default (a setting allows 1 to 8), and the map shows the next 28 days unless you choose other dates. Weather inside that window is as good as the public forecast. Third, Score and learn counts how many items have had their week close and how often planners marked one as new to them. For a chain near 2,000 stores, the per-store lookups run as a separate job ahead of the review. It is not a forecast and does not feed one.

What is the Watch list? Available today

Watch is one page with four tabs: Map, List, Alerts and Impact Memory. The List tab (the Forward Radar of earlier releases) holds what the review of the chosen week had on file near your stores for the weeks after it: road work, large local events, weather forecasts and alerts, competitor openings and closings, store operations and other records you supply. Each item names the store it is near, how far away it is, the dates the record covers and where the record came from. When Kelynto has collected the forecast behind an item, the item also shows when its source issued it and how it has changed since. An item is context for the planner, like the records on file in the weekly review. Planners mark each item as new to me, already knew or not relevant. The list needs store coordinates. The Map tab shows what is on file today for the days ahead.

Can we export everything Kelynto holds for us? Available today

Yes. A workspace admin starts a full export in the Admin pages and downloads one archive when it is ready. It holds every table of the workspace as JSON Lines, the stores and weekly figures also as CSV for a spreadsheet, the audit log with the hashes that let it be verified outside the product, the loaded files that are still kept, and a manifest that lists every file with its row count, size and SHA-256. A checksum file lets the whole archive be checked with a standard tool. Secrets such as key hashes are left out, and the manifest says which. The archive is kept for seven days by default and can be removed sooner. Members' sign-in accounts belong to the people, not the workspace, so they stay outside it.

How do you stop the commentary containing wrong numbers? Available today

Every number in a draft is parsed (money, percentages, counts, dates, scale words, numbers as words) and must match a number in the week's facts within the rounding its format implies. One unmatched number, or a wrong currency symbol, fails the draft; the template writer is used instead and the reason is shown on the run. The guard checks numbers and currency symbols; direction words such as above or below, and fractions written as words such as half, are read by the planner. A planner reads and publishes every review, and their edits are checked on every save: an amount or percentage that matches no figure of the week is named on the review and before publishing. The check reads figures, and the planner reads which cause or category a figure is written beside. The causes come from the rules engine, never from a language model.

Is there documentation and help inside the product? Available today

Yes. Every page of the product has a Help button that lists the tasks of that page with their steps, and a search that finds a task on any page. A short tour introduces the pages a person's role can open; it can be left at any stop, resumed later and started again. The documentation is a separate site written around tasks: getting started, the data contract and import templates, the Monday review, Compare, Watch, alerts and Impact Memory, causes added by planners, Score and learn, licensing, sending and follow-up, administration, security controls, an API reference generated from the product itself, and release notes. It is a folder of static pages with its own search, delivered with each deployment, and the Help panel links into it from wherever the operator serves it.

Does a person review the explanations before they go to leadership? Available today

Yes, always. Kelynto drafts; a planner decides. Each finding can be confirmed, rejected or corrected to a different cause, and the commentary is only published when a user with the planner, analyst or admin role publishes it. Nothing is sent to leadership automatically, and a published review can be reopened. Those verdicts are recorded as labels: they feed Score and learn and can be exported by you. Every edit, label and publish action is written to the audit log with the person who did it. If your planners reject most causes, Score and learn will show it plainly, which is the point.

What if Kelynto gets a cause wrong? Available today

It will be wrong sometimes, and the product is built around that. A cause in the draft is a candidate shown with the evidence on file for it, not a verdict. The planner rejects or corrects it before anything is published, and nothing reaches leadership without a person publishing it. Each rejection is kept as a label and shows up in planner agreement and the correction and rejection rates, so a pattern of wrong causes is visible within weeks. Where the evidence is thinner, the draft words the claim cautiously. Variance with no supported cause is left as not yet explained. The number check covers figures, and the planner reads direction words and names. The planner remains the author of the review; Kelynto is the first draft.

What is Impact Memory? Available today

Impact Memory is a searchable record, by store and week, of what was on file (weather, events, road work, competitor openings, closures, promotions, low in-stock) beside what sales and plan were. For a store and a kind of condition it lists the earlier times that condition was on file there, newest first, each with actual and plan for the weeks around it, what else was on file, and what the planner recorded then. It is a record and not a cause: it does not say a condition produced a result, it computes no average effect, and it forecasts nothing. A few occurrences are shown as a few, one by one. Each condition says when it came on file, and the history can be read as it stood on an earlier day.

What does the in-app assistant not do? Available today

It does not guess. It answers a fixed set of kinds of question, each through the same reading the screens use, and when nothing is on record it says so and says what would be needed. Every amount, percentage, count and date in an answer has to be on a record that answer cites: a checker runs on every answer before it is shown. A sentence that fails is left out, and the answer says so; if that is the sentence answering the question, nothing is given. A weather, event, road work or competitor record is listed as context and is never given an amount or called a cause. It shows no percentage of how sure it is. It forecasts nothing, gives no advice, writes no text for you and changes nothing: it only reads. The record behind each sentence is one click away, so you can check any answer yourself.

Can I ask questions inside the product? Available today

Yes. Every page of the product has an assistant panel, opened from the top bar or with a keyboard shortcut. You ask in plain words and it answers from your workspace's own records: what explains a gap at a store, region or category, which stores missed plan most, what changed against another week, what is on file around a store, which follow-up actions are open or overdue, what planners decided, how past explanations held up, what data is loaded, and what a term means. Each sentence carries a numbered source, and the number opens the record it came from. It knows which page, week and level you are looking at and says so. When a name fits more than one store it asks which one you mean. Your conversations are your own: you can reopen, rename and delete them.

What is the Experiments page? Available today

Experiments (named Investigations in earlier releases) estimates what a campaign changed, from a randomised study you supply as a JSON file: the study's design, its outcomes and one record for every assigned unit, including those who did not buy. It checks the study first. When the checks pass, it estimates each campaign group's effect per assigned unit against the control group, with an interval, and saves a draft brief with the evidence and what the analysis does not establish. Otherwise it says what evidence is missing.

A planner, analyst or admin runs one; anyone who can read reviews can reopen a saved one. It trusts what the file says about the design, and it does not explain a week's variance to plan: the weekly review does that.

Does Kelynto learn from our planners? Available today

It learns in a narrow, transparent way. Each confirmation or rejection is kept as a label on that type of cause in your workspace, and later drafts word that type of cause more or less cautiously as the labels add up. Edits feed the edit rate, marks on Watch items feed the novelty rate, and pulse answers feed the would-use score. Attribution is rules rather than a trained model, chosen because ground truth comes from planners labelling real weeks. The labels are also exportable as a dataset (signal, business impact, human-confirmed cause, outcome) that belongs to you. Kelynto does not use your data to train any model. We make no claim about how quickly results improve; Score and learn shows it on your own weeks.

What does a planner's Monday look like with Kelynto? Available today

The closed week's data arrives by upload or scheduled export. A scheduler produces the review at the hour you set. When the planner opens Review, the draft is written: the headline variance, a line that sets this week beside the week before measure by measure, and the week in three parts for the company and each level. The planner confirms, rejects or corrects each recorded cause, edits the wording and publishes. Under the three parts, "What this means for next week" says in words what carries into next week by the records (a stockout still open, a promotion that continues or ends), what stays not explained as an open question, and the open follow-ups, with no forecast. A Monday digest can email each member the published review of their own scope.

What is the noise guard? Available today

Store and category numbers wobble every week for no reason. The noise guard stops that wobble being reported as a finding. A store and category move with no known cause is listed as not yet explained only if it clears your two materiality bars (a percentage and a minimum amount) and is also larger than a multiple of that category's ordinary spread between stores that week. The default multiple is 3; lowering it shows more cells, and 0 switches the guard off. The minimum amount is sized to your business from your own sales when they are first loaded, and an admin can change it. The spread is measured from your own data each week, so it needs no tuning to start. In local testing on data with no real effect, a fixed 4 percent bar flagged about one cell in fifteen; with the guard, a handful. Named causes must pass a separate statistical gate.

Does Kelynto explain variance at product (SKU) level? Not available

Not today. The lowest level of the review is the category within a store and week. The engine that estimates causes works at that level and provides no product-level output.

What exists: a sales file may carry a product column. Kelynto adds the product rows up to category for the review, exactly as before, and keeps the product rows as you sent them. The Data page shows how many products are on file each week and what share of the week's sales they cover. On a category, the review says that product sales are on file and that category is its lowest level.

Kelynto does not work out product-level causes by dividing a category's estimate among its products. A product level will appear only when the engine provides one.

What user roles does Kelynto have? Available today

Four roles. Viewer reads published and in-progress reviews; it suits leadership and store operations. Planner edits the commentary, confirms or corrects causes and publishes. Analyst does everything a planner can, plus data uploads, re-runs and dataset exports. Admin does everything, plus users, workspace settings and the audit log. Permissions are checked on every API route, not only hidden in the interface. With single sign-on, roles come from your identity provider, so removing a role there removes access on the user's next request. An admin's Roles page lists what each role can do, taken from the rules the product enforces, and what one person can do and why. An admin can also limit a member who is not an admin to the stores of some regions, districts or stores: that member reads those stores only, with no company total, and does not edit the commentary, label causes or publish.

When does Score and learn show a result instead of insufficient data? Available today

Each measure stays at insufficient data until it has enough evidence: 10 labels for planner agreement, 3 pulse answers for would-use, and 2 reviews with planner edits for the edit rate (which is then the median of up to the last four). The overall verdict is shown only when at least two measures have enough data, and it says the measures meet their targets only if every judged measure does. Supporting figures are shown beside the three: how many Watch items have had their week close with sales on record (the same store, record and week listed by several runs counts once), how often Watch records were new to the planner, and planner agreement by cause type.

What does Score and learn measure? Available today

The page is called Score and learn. Three of its measures have a default target that you can change per workspace.

Draft edit rate: how much of the draft the planner changed, median of the last four reviews. Target 20 percent or less.

Planner agreement: the share of planners' labels that confirmed what the review recorded or had on file. Target 60 percent or more. It is agreement with the review, not ground truth, and is never called accuracy.

Would use every week: the planner's weekly answer on a 1 to 5 scale. Target 4 or higher.

Also shown, with no target: correction and rejection rates, coverage, time to publish, commentary edit distance and causes planners added.

Everything is computed from what planners actually did, never from synthetic data, and a sample workspace scores nothing. These are targets for a pilot, not results.

Can Kelynto send the review to leadership? Available today

Yes. A planner, analyst or admin sends a review from its Send button: a short summary, an optional note, and the review as a Word, PDF or PowerPoint file, a link, or both. Each person gets the review they may read. It goes to members of your workspace and to addresses at email domains your admin approved, and to nobody else. For people who never log in, the Monday pack mails the week's published review at the day and time your admin sets, Monday at 6:00 AM by default, to the addresses they name. Each address confirms first, and every email has a one-click unsubscribe and a link that expires, where the review reads online or downloads as PowerPoint, Word or PDF in your branding, with no account. Your admin chooses whether the files are also attached. With no mail relay configured, Kelynto sends nothing and says so. Every send is in the audit log.

How much time does Kelynto save? Not available

There is no measured figure yet, so none is quoted, and any hours saved or return on investment number would be a guess. The product measures it for you: at the start you state how long the Monday review takes today, and each week the planner records the minutes the review took. After a few weeks you set your own two numbers side by side. A useful first step is to tell us who writes the review today and how long it takes.

How are time zones and fiscal weeks handled? Available today

Both are workspace settings. The fiscal week can start on any day; files can carry a week start, a week ending date or daily dates, and daily rows are summed into your fiscal weeks. Set the week start before the first load, because rows are aligned as they are read. Signals are matched to the stores' local week: a week runs from local midnight to local midnight in your time zone, including across daylight-saving changes, so a Sunday night event counts in the week it happened. The limit: one time zone per workspace, so a chain spanning several zones should choose the zone most of its stores are in. Reviews are weekly; there is no fiscal period or 4-5-4 month roll-up today.

What happens when Kelynto cannot explain a variance? Available today

It says so, with a number. Every review states the amount not explained: every shortfall, store by category, that no cause your own records show is credited with. A lift elsewhere never covers it, so a promotion cannot hide a drop at other stores. It is on screen at every level, for the company, a region, a store and a category, whether it is small or nearly everything. A shortfall is never assigned to the nearest plausible story: a flood warning or a road closure near a store is listed as context on file, and is not credited with an amount. Where a promotion is on the calendar and no estimate can be stated for it, the review says that too. How much is unexplained depends on your data, and the review shows it every week. The unexplained part is useful: a planner can name the cause they know, and that is kept as their judgement.

How does Kelynto break down variance to plan? Available today

Variance is actual sales minus plan sales, computed for every store and category you send and rolled up to category, region and total. The rollups reconcile to the total, and that is enforced by tests. In the app the planner can view the week by region, by category or by store, and open a single store to see its categories and the signals near it. A week in which gains and misses nearly cancel is described in amounts above and below plan instead of shares, because a share of a near-zero net is meaningless. The grain is store by week by category: item-level data is not required, and item-level explanations are not produced.

Can Kelynto alert us about upcoming events near our stores? Available today

Yes. A person makes alert rules for themselves, and an admin can make rules for the whole workspace: which kinds of record, near which stores or regions, how many days ahead. You are told in the app, and by email if you choose, when such a record comes on file. If it later changes or is withdrawn you get a follow-up that says what changed. Email can arrive as each record does or as one daily digest, quiet hours hold email back, one message carries everything waiting, and each rule can be unsubscribed from. Email needs a mail relay configured for the deployment; without one each message is drafted and shown as not sent. An alert states the record and its source. It does not say what the record will do to sales.

What does the Watch map show? Available today

The Watch page shows what is on file near your stores for the days ahead on a map and a list that stay in step: weather, local events, road work and competitor openings, each with its place, its dates, its source and when the source issued it. You filter by kind, dates, region, store and distance, and stores close together are drawn as one mark until you zoom in. The map needs no map service and no key. It uses an open-source map library and draws over outlines the app carries, so by default the browser contacts nothing outside your deployment. An operator can configure a tile source to put a street map behind the stores. Everything on the map is also in the list, which is how a keyboard or a screen reader goes through it. A record is context: the page does not say what it will do to sales.

How do we record competitor openings and local events? Available today

There is no free public feed of competitor openings or local events, so your team records what it knows. A planner, analyst or admin uses a form on the Watch page; many records at once go in a signals file. A record needs a place (a store, or coordinates), dates and where you learned of it, and has no field for an expected effect. Records are kept as issued, like every feed: correcting one writes a new version and the earlier one stays on file with the time it was stated, and withdrawing one marks it withdrawn and deletes nothing. The record appears on the map and the list and in the weekly review as a record on file. A paid events source can be connected, and stays off until a key is configured for it.

Does Kelynto show what happened after a review or a decision? Available today

It shows what happened, and it does not judge. For each reviewed week, Score and learn lists what the review said, what planners decided for each item, and the plan and actual sales on file for the same stores and category in the following weeks.

This quarter with Kelynto does the same for a period of whole weeks: the weeks reviewed and published, what the reviews explained, and every decision recorded with its owner, date and status, beside what followed where the product records it (the Outlook's look-back, or Impact Memory's actual and plan). It downloads as a PowerPoint deck or a Word document, every amount checked against the stored records.

Neither page says an explanation or a decision was right, and neither states a return or a saving. Two data checks set a weather forecast as issued beside what its source later recorded, and a recorded promotion beside the calendar. Planner agreement is never called accuracy.

What is Kelynto? Available today

Kelynto is variance commentary for retail planning: the weekly review, written from your own records. It writes the review of sales against plan by store and category, with the evidence behind every cause, shows what is on file near your stores and what is coming against plan, records what planners decided, and counts how the reviews served them. It works alongside your planning system.

In practice, it writes the weekly business review that a planning team otherwise assembles by hand. A cause gets an amount only when your own records show it, and what the records do not explain is stated as an amount. Retail planning teams use it as their own application, and planning services, software vendors and consultancies offer it to their clients under their own brand. The way to judge it is a pilot on your own past weeks, which starts with a 20-minute conversation.

What does Kelynto deliberately not do? Available today

It does not replace your demand forecast, build the plan, replenish, allocate, price or order. It is not a BI or dashboard tool, not a supply-chain suite, and not a general chatbot: the assistant inside it answers a fixed set of questions from your own records, with the record behind each sentence. It does not write back to any planning system or ERP. It does not need or hold transactions, baskets, shopper, payment or employee data. Watch lists records near your stores with no amount; the Outlook sizes what is coming against your plan or forecast. Attribution is rules, not a trained model. The job is narrow on purpose: explain last week's variance to the plan you already have, show what is coming near stores, capture what the planner confirmed, and score whether that was useful. If you need a new forecast engine, keep looking; if your team writes a Monday review by hand, this is the part it covers.

What does Kelynto do? Available today

Kelynto drafts the Monday business review for a retail planning team. Each week it takes sales against plan by store and category, plus in-stock, promotions and local signals, and drafts the review in three parts: what your own records show with an estimated impact, what is on file near your stores, and how much is not explained. The planner edits the draft, confirms or corrects what is on record, and publishes. It also shows what is coming near your stores in the next few weeks, and counts, from real use on its Score and learn page, whether its explanations are proving useful. It sits above the planning system you already use and replaces none of it. The quickest way to judge it is a 20-minute conversation about your own Monday review.

Who is Kelynto for? Available today

Two kinds of company. Retailers' planning teams, who owe leadership a weekly explanation of sales against plan: demand planners, merchandise planners and analysts use it, and directors and vice presidents in planning, merchandising, supply chain and store operations read it. And the companies that serve retailers: planning and open-to-buy services, retail software vendors and consultancies, who offer the review to their clients through the partner API, embedded views and their own brand, one workspace per client. It fits wherever sales are reviewed against a plan by store and category, from a single store working with a planning service to chains of several hundred locations; larger chains start with a sizing check on their own volume. The explanation draws on stockouts and promotions in your own files, and on weather, local events, road work and competitor openings near your stores.

Partners

Can a software vendor or partner offer Kelynto inside its own product? Available today

Yes. A partner's software calls the partner API with a partner key, and each retail client has a Kelynto workspace of its own. The key can create a client workspace, send that client's files through the same validation as every upload, produce and publish its weekly review, read the review, findings and Outlook, download it as Word, PDF or PowerPoint, and manage the client's signed webhooks. Every call names one client and is bound to that one workspace; a client that is not the partner's answers not found. Keys are scoped, rotated and revoked, and every request is in the client's audit log. The partner's own pages can also show a client's review, Outlook and This quarter in the partner's brand, through single-use embed links only the partner's site can frame. Partner admins have a console; developers start from a sandbox of synthetic clients and the documentation at https://kelynto.com/developers/. Kelynto also stays a standalone application for each client's own people.

How are a partner's clients licensed? Available today

As a pool. With a wholesale agreement a partner holds a pool of client licences. Each client workspace the partner makes takes one, on the agreement's terms and at the tier the partner asks for, and gives it back when the client leaves the partner. The partner console and the partner API show the pool's size, how many licences are allocated and available, each client's tier and end date, and a warning when the pool is nearly full or the agreement ends soon; a webhook can carry the same warnings. A client can be given another tier the agreement allows. The product holds counts and terms only: what a partner pays is in its signed agreement, and is agreed in a conversation with the founder.

Who owns the client relationship and the client's data, and can a client leave its partner? Available today

The relationship stays with the channel that holds it, and the data stays the client's. Every client workspace records one channel, Kelynto directly or one partner, with who originated it and its history, so a partner's clients are the partner's on the record. Each client's data and the reviews produced for it belong to the client, and the partner's export records each client's consent before that client's data loads. A partner's key publishes to a client's readers only when that client's admin has allowed it, and every partner request is in the client's audit log. A client can move to Kelynto directly or to another partner with its data in place, with both sides' consent on record, and the former partner's access ends at that moment.

Can one export from our system carry all our clients? Available today

Yes. A partner's export can carry many clients in one file, with a manifest that names the batch, the row count, each client's consent, its currency, its week convention and its control totals. Each client is loaded into its own workspace through the same validator, delivery gate and read-back as any upload, and a client's sales must reconcile to its control totals to the cent or that client's load is held for a person to decide. A client whose consent is not recorded is not loaded. A dry run checks a file and writes nothing. The same clients can also be fed one at a time through the partner API, where every call names one client and is bound to that one workspace.

How do we get a sandbox and the API documentation? Available today

The developer documentation is public at https://kelynto.com/developers/: an overview, a quickstart in curl and Python from a key to a published review, the embedding guide, every route with its scope and error codes, and a Postman collection. For a sandbox, write to marc@kelynto.com with the subject Sandbox access, your company, your name and role, and the origin you will embed from if you plan to embed. You receive a base address and a sandbox key that reaches two synthetic client workspaces, and the synthetic clients the key makes itself, for 30 days. The key holds every scope, so a developer can send a file, produce and publish a review, download the documents, add webhooks and frame the embedded views. Every name in the sandbox is marked synthetic.

How many clients can one partner run on Kelynto? Available today

It was measured at 200 client workspaces, on synthetic data. On one 2-core machine with PostgreSQL on the same machine, a partner export of 200 clients (492 stores, 91,440 rows) was imported in about three and a half minutes, each client into its own workspace, and the weekly reviews of all 200 ran in about 32 seconds, four at a time. The partner view of all clients answered in a median of 26 milliseconds. The weekly run is one job per workspace, a failure in one blocks no other, and a partner's run summary says which clients succeeded. These figures say how the product behaves on data of that shape and size; a partner with more clients or larger ones starts with a sizing check on its own volume.

Is Kelynto a fit for our clients if they have one to five stores? Available today

Yes, and most often through the planning service or software company that already serves them. A workspace can hold a single store. Kelynto reads the standard exports of the point-of-sale systems independents run (Lightspeed Retail, Square for Retail, Shopify POS, NCR Counterpoint, Retail Pro and others) and an open-to-buy plan by month and class, so a small retailer's sales and plan arrive as the files it already has. For a planning service the gain is on the advisor's side: every client's week explained before the advisor opens the file, on one page ranked across all clients, and a write-up to send under the service's own name. Each client has a workspace of its own. A retailer of any size can also come to Kelynto directly.

How can a planning service, a software vendor or a consultancy work with Kelynto? Available today

In four ways, chosen by who sells and who supports. Refer: you introduce the retailers who ask you, and Kelynto serves them directly. Resell: you hold the relationship and the contract, and Kelynto runs the service for the retailers you serve. Embed: Kelynto sits inside your own product, in your brand, through the partner API and embedded views. Implement: you set up and run Kelynto for the retailers you advise, alongside their planning system. In every one, each client has a workspace of its own on Kelynto Managed, and the client's data and reviews stay the client's. For a partner, Kelynto is an explanation layer on top of your planning service or product. Terms are agreed with each partner in a conversation: book a 30-minute partner conversation from https://kelynto.com/partners/ or write to marc@kelynto.com.

Can we offer Kelynto to our clients under our own brand? Available today

Yes. A partner's brand goes on what its clients read. Once your brand is recorded, the embedded views carry your name, your logo and your primary colour, and so do the documents Kelynto produces for the retailers you serve: the review as Word, PDF and PowerPoint, the emails a review is sent with and the Monday pack. The mail can carry your name as its sender, and can come from your own domain once that domain is verified. Embedded, your people read a client's review, Outlook, causes and This quarter inside your own application, with no second sign-in. Each client workspace also stays a standalone workspace its own people can open. To see it on synthetic clients, book a 30-minute partner conversation from https://kelynto.com/partners/.

Pilot

What happens after the pilot? Available through enterprise implementation

A readout of the three measures and a decision. The pilot runbook reads the results this way: all three meeting their targets means the product fits this team, and the next step is to continue and automate the data feed. A high edit rate with good planner agreement means planners accept the causes and rewrite the wording, so the writer is tuned. Low planner agreement means causes are wrong or missing: read the corrections before continuing. A low would-use score despite good planner agreement means the pain is not strong enough here, and we would say so. Commercial terms for continuing are agreed in a conversation with the founder. If you stop, you keep your exports and your data is deleted on the agreed date.

What does a pilot cost? Planned

Pricing is not finalised. It will depend on deployment model, number of locations and pilot scope. The fastest way to get a number for your situation is a 20-minute conversation.

The fee for a pilot is agreed per pilot, in that conversation. The paper is agreed there too: a mutual NDA before anything is shared, then a pilot agreement and a data processing addendum before the first upload, and for the customer-hosted option a statement of how any access by Kelynto is granted. Drafts of each are ready for review by counsel, and whether Kelynto signs your own paper instead is a founder decision. A pilot asks of you weekly files, a custodian for the sealed weeks, two planners for a few hours of judging, and a planner's attention each live week. Book 20 minutes at https://calendly.com/kelynto/30min or write to marc@kelynto.com. This is not legal advice.

What do we have to commit to a pilot? Available through enterprise implementation

Four named people. A sponsor who signs the pilot, names the others and receives the readout. An IT contact who registers the application in your identity provider, assigns roles and handles the security review. A data contact, usually an analyst, who produces the files, reads the validation reports and owns the weekly export. And a planner who reviews each week's draft, confirms or corrects the causes and publishes. The planner's weekly part is the same work they do today, starting from a draft, and the product records the minutes each review takes, so you see the real figure for your team. The steady commitment on your side: a file arrives every week and a planner engages. Your current process keeps running throughout the pilot.

How is a pilot evaluated? Available through enterprise implementation

In two ways, both on your data. Before going live, the eight sealed weeks: a custodian on your side holds your last eight weeks and the commentary your team actually sent until our set-up is frozen, we run them once, and two of your planners judge samples without seeing our output. The scorecard's bars are fixed before the seal opens and a miss is reported as a miss. Then the live weeks feed the Score and learn page, which is counted only from what your planner did: how much of the draft they changed, which causes they confirmed, and whether they would use it every week. Each measure shows insufficient data until it has a minimum sample. Beyond the numbers, two things are worth watching: whether the planner is editing or rewriting, and how often the planner marks a Watch record as new to them.

How do we exit a pilot, and what happens to our data? Available through enterprise implementation

Offboarding is a documented procedure. You agree an end date and a deletion date in writing. Before the end date you can export your labeled findings, Watch outcomes, published commentary and audit log. On the end date access is blocked and the data kept, which can be undone. On the deletion date one operator command deletes the workspace and every row it owns in a single transaction, verifies nothing is left, and prints a deletion record that we send you. Two things remain for a while, stated up front: database backups hold the deleted rows until backup retention passes (35 days in the production settings), and logs hold request metadata, never sales figures, until log retention passes. The exit terms themselves belong in the pilot agreement. The deletion procedure is rehearsed end to end.

How do we start a pilot conversation? Available today

Book a 20-minute conversation with the founder at https://calendly.com/kelynto/30min, or write to marc@kelynto.com. It is a working conversation, not a pitch: how your Monday review gets written today, who writes it, what data you plan with, and whether a pilot would answer a real question for you. Useful things to bring: your number of stores, the planning system or spreadsheet your plan comes from, whether you plan at store and category level, and one past commentary if you can share it. If it looks like a fit, the next steps are a short pilot agreement, the data and identity checklist, and a backtest on your past weeks. If it is not a fit, we will say so.

What if we do not have single sign-on? Available through enterprise implementation

Two routes exist in the product. First, a shared sign-in provider: one identity provider that Kelynto sets up for customers who cannot register an application in their own directory for a pilot; a signed claim in the token selects your workspace and roles are stored on the account in Kelynto. Your users' credentials and multi-factor settings are then held by that provider, which becomes a subprocessor. Second, the product's own sign-in by emailed single-use link or code, with no password, which the operator switches on; a person can add an authenticator app as a second factor. In both, revoking a user means deactivating the account in Admin, and you can move to your own directory later with accounts keeping their history. Both were rehearsed end to end with a test identity provider and a test mail server, and the provider and mail service for your pilot are set up during onboarding.

What are the onboarding steps? Available through enterprise implementation

In order: a kickoff call to name the people and agree the fiscal week start, time zone and data timing. A Kelynto operator creates your workspace. Your IT contact registers the application in your identity provider and assigns roles, or you use the shared sign-in provider if you have no single sign-on. Sign-in is proven on a call with an admin, a second role and a user without a role. Your data contact exports the files; they are loaded and every validation report is read together. A readiness check runs, the first week is run by hand and read by the operator, and your planner is walked through the first draft. Then the weekly export and schedule are switched on. These steps were rehearsed end to end with a synthetic retailer.

What is a Kelynto pilot? Available through enterprise implementation

A pilot answers one question on your own data: would this planning team use Kelynto every Monday? It starts with a backtest we call the eight sealed weeks. You send 26 weeks or more of weekly sales and plan by store and category, your store list, and in-stock and promotions if you have them. A custodian on your side, outside the planning team, holds the last eight weeks until our set-up is frozen and recorded. We then run those eight weeks once and give you each week restated in three parts, a scorecard against bars fixed before anyone saw a result, and a written finding whichever way it goes. If it earns it, about eight live weeks follow, with one planner editing, confirming or correcting each week's draft. The measures are fixed in advance so the result is judged in the open.

What is a proof run, and what is in the proof pack? Available through enterprise implementation

A proof run is Kelynto run over your own past weeks. You send an export of your history: fifty-two weeks of sales by store and category with the plan as it was approved, and in-stock, promotions and later forecasts where you have them. For each week, the Outlook is issued as of that week for the weeks after it, reading nothing later, and the weekly review of that week is produced the ordinary way. The proof pack is what comes back: a Word document and a PowerPoint deck that state, week by week, what the review would have said, whether the Outlook's range held for the weeks ahead and whether its error against what was sold was lower than the plan's, in counts and sums, and two CSV files with every row behind them. Every page names your workspace, the weeks, the release and the engine version. The workspace made for the run is deleted afterwards unless you go ahead.

What data do we need to provide for a pilot? Available through enterprise implementation

Five things; your own column names are fine. A store list with store code, region and coordinates for all or nearly all stores. Weekly sales and plan by store and category, the plan as it was approved: 26 weeks or more, 34 or more if you can. In-stock by store, week and category if you measure it, or an acceptance that stockouts will not be explained. A promotion calendar and a list of closures, remodels and competitor openings if you have them. And the commentary your team actually sent for the last eight weeks, held by a custodian on your side with those eight weeks of data until our set-up is frozen. Two facts before the first load: the day your fiscal week starts and the time zone most stores are in. CSV or Excel, aggregates only: no transaction, shopper or employee data. If your plan comes from a named planning system, we have a file recipe for its export.

Do our planners have to log in from day one? Available through enterprise implementation

No. The pilot runbook allows the first weeks to be run by hand: you send the week's files, a Kelynto operator runs the week, reads the draft and evidence, and sends the exported commentary by email. Your planner replies with edits and verdicts on each cause, and the operator enters them on the planner's behalf. The edit rate and planner agreement are the same either way. This keeps the pilot moving while your IT team registers the application in your identity provider, and it can be a smaller thing for IT to review at the start. The trade-off is that in this mode Kelynto staff handle your files and enter your planner's feedback, so it should be covered in the pilot agreement.

What are the success criteria for a pilot? Available through enterprise implementation

Three default targets, agreed before the first live week and adjustable for your workspace. Draft edit rate at or below 20 percent (median of the last four reviews). Planner agreement at or above 60 percent (at least 10 labels): the share of labels that confirmed what the review recorded or had on file. It is agreement between a planner and the review, not a measure of who was right. Would use every week at 4 of 5 or higher (at least 3 answers). These are evaluation criteria, not a performance guarantee. If your own bar is different, for example a specific question leadership keeps asking, say so at the start so it can be written into the pilot agreement.

How long does it take to get started? Available through enterprise implementation

These are planning estimates. The backtest, the eight sealed weeks, is planned at ten business days from the day your files arrive: files loaded and read with your data contact on day one, set-up frozen by day three, your custodian releases the eight weeks on day four, two of your planners judge samples on days five to nine, and the written result arrives on day ten. The pace depends on your planners' time, about two to three hours each, and on whether the first files load clean. If the software runs in your own cloud subscription, your platform team's set-up comes before day one and its length depends on your team. Live weeks, if you choose them, run about eight weeks. Your planner can start without signing in on day one.

Data

Is there a screen to map our columns ourselves? Available today

Yes. In the web app a file is checked and shown as a preview before anything is loaded. Column names the data contract knows are mapped by themselves. For any other column the preview proposes a field from the name and the values, and you confirm or change each one; nothing is applied until someone confirms it, and a confirmed mapping is remembered for next week's file, however it arrives. Repairs with one possible reading are applied and reported. Some things are never guessed, such as whether 03/04 is the third of April or the fourth of March: you are asked once. A confirmed mapping can be kept as a source profile from the same screen. A recap workbook laid out as a grid is read the same way.

Do we have to rename our columns to match your format? Available today

Usually not. The validator recognises many common names for each field, for example Store Nbr, Dept, Net Sales, Forecast, LY Sales and OSA, and the upload report shows exactly how each of your columns was mapped. It also reads real-world number formats: currency symbols, thousands separators, negatives in brackets and percentages, and it detects whether in-stock is on a 0 to 1 or 0 to 100 scale. If a required field is not found under any known name, the file is rejected and the report lists the columns it did not recognise; one of them is usually the missing field under another name. In the web app the upload preview then proposes which of your columns is which field; you confirm it once and it is remembered for the next file. Through the upload API alone, the fix is to rename the column. The full list of accepted names is shown on the Data page.

Which systems' exports does Kelynto read as they are? Available today

Kelynto reads the standard exports of Lightspeed Retail (R-Series and X-Series), Square for Retail, Shopify POS, NCR Counterpoint, Retail Pro and Microsoft Dynamics 365 Business Central (Item Ledger Entries), as well as an open-to-buy plan by month and class and a weekly recap with stores down the side and weeks across the top. On the first upload it knows the export by its columns, says which export it read the file as, and maps every column without a mapping screen. You can choose another export, or none, on the preview, and a saved source profile remembers the choice. Clover line items and RICS weekly sales are read with a ready-made mapping you pick on the preview. Each shape is built from the vendor's public documentation, and every file goes through the same validator as any upload.

What file formats are accepted? Available today

CSV, Excel .xlsx and the older Excel .xls. Delimited files may use comma, semicolon, tab or pipe, and the separator is detected. The encodings exports usually come in are read: UTF-8 with or without a byte order mark, Windows-1252, and UTF-16 (what Excel saves as Unicode Text). In a workbook with several sheets, the sheet whose columns match the dataset is read; when that cannot be told, you are asked which. A sheet laid out as a grid, with weeks or categories across the top, is read as rows. Not supported: password-protected workbooks, Parquet, JSON, XML, EDI and fixed-width files; a file that cannot be read is refused with what to change. Large files should be sent as CSV: a CSV is parsed from disk in blocks, while a workbook is unpacked in memory and loads at roughly half the speed. The default upload limit is 50 MB per file; a larger history can be split by date range.

We have daily or item-level data. Can Kelynto use it? Available today

The working grain is store by week by category. Daily rows are accepted and summed into your fiscal weeks. Rows below category level, such as a file split by sub-department or item, are added up to the category in the file, provided each row carries the store, week and category; the report warns how many rows were combined. What you call a category is your choice: department, category or subclass all work, as long as sales and plan use the same level. Explanations are produced at that level and no lower, so item-level questions (which SKU was out of stock) are outside what it answers. Sending aggregates is also the lighter option for your security review.

Our data is messy. Will that be a problem? Available today

Most mess is handled; a few things are fixed at the source. Handled: your own column names, currency symbols and separators, blank lines, a title above the header row, a grid with weeks or categories across the top, a totals row (checked against the rows and kept out of the load), mixed encodings, week-ending or daily dates, and placeholder dates (flagged as row errors). Names differing only in capitals, spaces or punctuation load as one name, listed in the report. A row that repeats an earlier row exactly is counted once. Rows for one store, week and category that differ only in amounts load after an analyst says whether they are one total or a record sent twice. Fixed at the source: store codes that do not match your stores file (those rows are skipped, and a close name is matched once you confirm it), a missing plan, and files where more than one row in five has an error.

How much history do you need? Available today

Technically one week of sales with plan is enough to write a review. Stockout attribution compares each store and category's in-stock with its typical level over the eight weeks before, so nine weeks of in-stock gives the full baseline. For a pilot we ask for 26 weeks or more of sales and plan. That lets past weeks be run and set beside the commentary your team actually wrote, and gives Watch outcomes something to be scored against. You do not need years of history, because nothing is being trained and no forecast is being fitted. If you have less than 26 weeks at store and category level, it is still worth a conversation about what a shorter backtest can show.

What if we do not plan at store and category level? Available today

The product explains variance to plan at store and category level, and the sales file carries a plan value on each row, so the review needs a plan at that level. Prior-year sales is an optional column; the review is built on the plan. Two good starting points: begin with the part of the business where a store and category plan exists, or talk to us about how your plan is spread to stores today; many teams have an allocated plan they do not think of as one. The Outlook's reference, the plan or forecast for the weeks ahead, can also be loaded at a coarser grain: by month or fiscal period and class, chain-wide if that is how it is planned. Kelynto spreads it to weeks and stores from your own sales history, records how, and uses it as the review's plan for a sales file that carries none. A 20-minute conversation settles this before any data is sent.

Does Kelynto need customer or personal data? Available today

No shopper, payment, transaction, basket or employee HR data is needed, and the data contract has no field for any of it. Please do not send it. What Kelynto holds is confidential business data, not consumer data: weekly sales and plan by store and category, in-stock rates, promotions, the store list with coordinates, and the commentary and labels your users write. The only personal data is business contact data about your users: work email, display name, role and last sign-in time, taken from your identity provider. No cardholder data, health information or government identifiers are processed, which is why frameworks specific to those data types are not in scope for the data itself.

Can we add our own local knowledge, like a competitor opening? Available today

Yes. A planner, analyst or admin can enter one on a form on the Watch page, and many at once go in the optional signals file. Each row is one local fact: a type (event, weather, road work, competitor opening, competitor closing, store operations such as a remodel, closure or outage, or other), a title, a start time, and either a store code or coordinates. You can add an end time, expected attendance and the categories affected. Those signals are then treated like the automatic ones: matched to stores by distance and week, listed in the review as records on file with their place and dates, and shown on Watch when they are in the coming weeks. A store closure can state the days the store did not trade: that week's review then estimates the closure from the store's plan for those days, where its sales show it. Files arrive by upload or the upload API.

Who owns the data and the labels our planners create? Available today

You do. The security overview states that the labeled dataset stays in your workspace and is exportable by you, and the pilot runbook states that the customer owns the exported findings and Watch outcomes. Analysts and admins can export the labeled findings and Watch outcomes as CSV and each review as a document at any time. Kelynto does not use your data to train any model, and workspaces are isolated from each other in the application and in the database, so one customer's data and labels are not visible to another. The precise ownership, licence and confidentiality wording belongs in your agreement; there is an outline of a pilot agreement, to be finalised with counsel, and that is the place to check it.

What data does Kelynto need? Available today

Two files are enough for a first review. Stores: one row per store with a store code and region; latitude and longitude are strongly recommended, because without them no weather, event or road-work signal can be matched. Sales: one row per store, week and category with actual sales and plan sales. Four optional files add depth: in-stock percentage by store, week and category (this is what makes stockout attribution possible, and is the single most valuable addition), promotions by week and category, receipts (what was planned to land in each store, week and category, what did, and when, in amounts or units, so that a late or short delivery is on file beside the week), and local signals only you know, such as a competitor opening or a remodel. Everything is weekly aggregates. Item-level, transaction, basket, shopper, payment and employee data are not needed and should not be sent.

What data is sent to the outside signal sources? Available today

Store coordinates and dates, nothing else. The United States National Weather Service receives a coordinate pair per store, for alerts and the point forecast, with a contact in the request header that you set to your own in a Customer Environment. Open-Meteo, only if an admin switched it on, receives coordinates rounded to four decimals and a date range. PredictHQ, only if enabled, receives coordinates, a radius and a date range. Road-work feeds receive nothing: the feed is downloaded and matched to stores inside the deployment. No store names, sales, plan, in-stock, commentary or user details go to any of them. A full set of coordinates does reveal your store footprint to those services, so each source has a switch per workspace, and one deployment-level setting forbids all of them for every workspace. Local facts can then be loaded as a signals file. The public sources are asked daily and at each weekly review; their answers stay inside the deployment.

What outside signal sources does Kelynto use? Available today

Five adapters exist. The US National Weather Service for alerts and point forecasts is on by default and covers the United States. Open-Meteo for daily weather values is off by default: its free service is for non-commercial use, so an agreement with Open-Meteo comes before a workspace switches it on. PredictHQ for local events is off by default and needs a PredictHQ key, a commercial service you or we would license. State road-work feeds in the WZDx format are used when an admin adds the feed addresses, up to 20. Each source has a switch per workspace, and one failing never blocks the weekly run. Foot traffic, mobile location, social media, competitor pricing and syndicated market data are not used. Every adapter has automated tests, and what the public sources answer is kept exactly as issued, from each daily collection.

What happens when we upload a file? Available today

Every file gets a report: accepted, partial or rejected, or, under a source profile, held. Accepted: every row loaded. Partial: some rows were skipped (unreadable values, or store codes not in your stores file) and the rest loaded. Rejected: a required column is missing, more than 20 percent of rows have an error, no row could be loaded, or a sales or in-stock file arrived before any store was loaded; nothing is loaded, so a broken export cannot overwrite good data. Every problem is located: the row, the column by its header and its spreadsheet letter, the value found, the rule it broke and a suggested fix. The complete list downloads as a file you can hand to whoever produces the export. Uploads are idempotent: a corrected file for the same weeks updates in place. A file loaded with a wrong mapping can be read again from its stored copy, with a before and after, and nothing is doubled.

Integrations

What would a connector to our planning system involve? Available through enterprise implementation

A connector is scoped implementation work, agreed for your systems. In most cases the weekly files are enough: the requirement is four weekly files at most, and a scheduled export from your planning system or warehouse to the upload API covers it. That work is usually a query, a schedule and a token from your identity provider, owned by your data contact, with the validator's report to confirm each load. A true connector, one that reads your system's own interfaces, means agreeing access, credentials, data mapping, error handling and who maintains it when your system is upgraded. The usual path is to start with the export, learn what your data looks like, and then decide whether a connector earns its cost.

Can Kelynto read directly from Snowflake or our data warehouse? Not available

No. Kelynto does not connect to Snowflake, BigQuery, Databricks, Synapse or any other warehouse, and it holds no credentials to your systems. Data arrives as files that your side sends: a scheduled job queries your warehouse, writes the weekly files in the documented contract and posts them to the upload API. For many IT teams that is the simpler pattern to approve, because data is pushed out by a job you control and nothing reaches in. The files are weekly aggregates by store and category, typically small. If a direct connection is a firm requirement, say so in a conversation so it is recorded as a real need.

Does Kelynto integrate with Blue Yonder, RELEX, o9, Oracle or Anaplan? Available through enterprise implementation

No built integration exists with any of them. Kelynto is independent of the planning system: it reads weekly files in a documented contract, so it works alongside any system that can export weekly sales and plan by store and category, including a spreadsheet. What Kelynto has is a file recipe for each common system (Blue Yonder, RELEX, Oracle Retail, SAP IBP, o9, Anaplan, Logility, and the weekly Excel recap): which export carries what, how that system defines the week, how promotions and availability appear, and the traps. Each recipe says what was confirmed against public documentation and what to confirm with your own administrator. Your team schedules the export and delivers the files by upload, API or a scheduled drop from a folder you name. Nothing is written back to the planning system.

Can we use Kelynto's results in Power BI? Available today

Yes, through a reporting export made for it: seven flat tables (reviews by week, the lines of each review, data coverage, data feeds, stores, the causes planners added and their decisions) as CSV and as Parquet, behind an authenticated endpoint, with a schema documented column by column that only ever gains columns. Download the files, or have a scheduled job or a Power BI scheduled refresh fetch them with an API key; a workspace admin makes the key on the Admin Security tab. The key needs the read data scope, and the read reviews scope as well for the four tables that carry what each review says. Load them with Power BI's Parquet or CSV connector. Every file is verified in automated tests by reading it back. An estimated amount appears for recorded causes only.

Can Kelynto work with SAP? Available through enterprise implementation

It works alongside SAP through weekly files, and there is no SAP connector or SAP certification. What it needs is a weekly export of sales against plan by store and category, a store list, and optionally in-stock and promotions. Any system that can produce those files can feed it, and there is a file recipe for SAP IBP. The usual pattern is a scheduled job on your side that writes the files and posts them to the upload API using a token from your identity provider. Your data contact sets up that export during implementation. A first evaluation can start with files exported by hand.

How do we automate the weekly data feed? Available today

Push: your scheduler posts each closed week's file to the inbound API with a key holding the write data scope. Pull: an admin sets up a scheduled drop on an SFTP folder, an S3-compatible bucket or an Azure Blob container, at public addresses. An analyst can attach a source profile, the export's versioned contract. A file whose columns differ, whose control totals do not reconcile to the cent, or that would change a published review is held with exactly what it would change, row by row, until an admin or analyst who did not send it approves it with a reason or rejects it. A resent file loads once and its control totals are still checked. Every load is read back, and every delivery is in a downloadable history. An approved correction is restated beside the published review. In your own cloud, drops stay off until your operator switches them on.

Can Kelynto pick up files from SFTP or cloud storage? Available today

Yes. A workspace admin adds a scheduled drop under Data: an SFTP folder, an S3-compatible bucket or an Azure Blob container, the dataset the files hold, a file-name pattern, how often to look (hourly to weekly), and whether a taken file is left, moved or deleted. Each file goes through the same validation as an upload, and every run is in a history with one line per file. Credentials are entered by your admin, stored encrypted, never shown again and never logged. An SFTP server's host key must be accepted before anything is sent to it. Drops reach public addresses; a server inside your own network pushes to the API instead. In a deployment in your own cloud, scheduled drops stay off until your operator switches them on. Drops are tested against stand-in SFTP, S3-compatible and Blob servers.

What integrations does Kelynto support today? Available today

Input is files in a documented contract: stores, weekly sales against plan, and optionally in-stock, promotions and local signals, as CSV or Excel. There are four ways to deliver them: upload in the app's Data page, send them to the inbound API, let Kelynto fetch them on a schedule from an SFTP folder, an S3-compatible bucket or an Azure Blob container that your admin names, or have an operator load a directory at the command line for the first load. All four go through the same validator and produce the same report. Nothing is installed on your side. Outbound, there is a read-only API, signed webhooks, and a flat reporting export as CSV and Parquet. It pulls weather, alerts, events and road-work feeds, and writes nothing back into your systems. Planning suites, ERPs and data warehouses connect through these files rather than built connectors.

Is there an API for sending data? Available today

Yes. The inbound API takes a file of any dataset: POST /api/v1/inbound/submissions/{kind}, where kind is stores, sales, inventory, promotions or signals, with the file as the request body. Further routes return a submission's status, its validation report and a file of every problem by row and column. An Idempotency-Key header makes a retry safe: the same key and file load nothing twice. The caller sends an access token for a user or service identity that holds the analyst or admin role in your identity provider, or a Kelynto API key with the write data scope. A workspace admin creates keys on the Admin Security tab; a key is shown once, stored hashed, expires within a year at most, and every use is in the audit log. The default size limit is 50 MB per file, and requests are rate limited. The credential, scope and limits are checked before the file body is read. The earlier upload endpoint still works.

Can Kelynto notify our systems when something happens? Available today

Yes. A workspace admin adds an https address and chooses the events: a file loaded or was held for a decision, a review was produced or published, an Outlook was issued, the Monday pack was sent, a scheduled drop failed, or expected data is late or missing. Each call is a small JSON message signed with HMAC-SHA256 using a secret that is shown once, so your receiver can verify it came from your workspace and is recent. A failed call is retried after 1, 5 and 30 minutes, then 2, 6 and 24 hours, and every attempt is in a delivery log; a test button sends one at once. A webhook points at a public https address, never at a private or internal one, and redirects are not followed. The message for a published review gives amounts for recorded causes only and no score. In a deployment in your own cloud, webhooks stay off until your operator switches them on.

Deployment

Can Kelynto run on AWS? Available through enterprise implementation

Available through enterprise implementation. A Terraform module for AWS (ECS on Fargate, RDS for PostgreSQL) is written and checked statically: against the AWS provider's published schema and with a policy scanner, with its scripts run against stand-ins. It mirrors the Azure kit that runs Kelynto Managed, with AWS services: Secrets Manager, an Application Load Balancer with AWS WAF, CloudWatch logs and alarms, your own VPC and your own KMS key if you want them, and scripts for preflight, deploy, validate, upgrade and rollback. The first install in your AWS account is planned and applied with your platform team as its validation run. If AWS is your requirement, say so early in a conversation.

Can Kelynto run in our Azure environment? Available through enterprise implementation

Available through enterprise implementation. The Azure deployment kit is the one that runs Kelynto Managed; in your subscription it is set up during the implementation. What the kit is: Bicep templates for a single-tenant install in your own subscription (Container Apps for the API, web app and scheduler, PostgreSQL Flexible Server on a private network, Key Vault, Log Analytics, optional Front Door), signing in with your own directory, plus scripts for preflight, deploy, validation, upgrade, rollback and uninstall. A private shape adds your existing virtual network, private endpoints, internal ingress and a database key you hold. The templates compile and lint, and the same container stack ran in production mode with a full onboarding rehearsal. The first deployment in your subscription is run with your team as its validation run. You own the subscription, network, keys and backups.

What data leaves our environment if we host Kelynto ourselves? Available through enterprise implementation

Kelynto can run inside your own cloud account. In that model your sales, plan and in-stock data, your store list and your users' details are stored and processed in your environment and are not sent to Kelynto. The software does not call home.

Your data does not have to leave your environment. By default Kelynto looks up weather alerts and forecasts for your stores by sending their coordinates (no names, no sales) to the United States National Weather Service. You can switch that off, for one workspace or for the whole deployment. Webhooks and scheduled file drops are off in this model until your operator switches them on. With weather off and the built-in writer, the only outbound connection is to your own identity provider, to download its public signing keys.

Available through enterprise implementation, with the Azure deployment kit that runs Kelynto Managed.

Can Kelynto run on Google Cloud? Not available

No, Google Cloud is not a supported target today. The deployment kits are for Azure, which runs Kelynto Managed, and AWS, as a Terraform module. For Google Cloud there is a one-page mapping of the components to its services, as a design statement. The application itself is three containers and a PostgreSQL database, and a Docker Compose path for any container host with your own PostgreSQL server runs the stack in production mode, so the application code depends on no particular cloud. If Google Cloud is your only option, the paths are the Kelynto Managed model, or a conversation about whether a container-based install on a virtual machine in your project would meet your requirements.

Can Kelynto run on-premises or on Kubernetes? Available through enterprise implementation

There is a container path rather than a packaged on-premises product. A Docker Compose file runs the whole stack (API, scheduler, web, PostgreSQL) in single-tenant mode, and an overlay points it at a PostgreSQL server you already run. Both were built and run in production mode, the second against a separate database server over TLS, with the same post-deployment validation the cloud kits use. What is scripted around the single-machine variant (virtual machine creation, TLS certificates, encrypted backups) assumes an Azure virtual machine. On your own hosts you supply the machine, TLS, backups, monitoring and outbound access to your identity provider's key endpoint. The kit has no Helm chart or Kubernetes manifests today. An install in your own data centre is run as an implementation project with your team.

Does the deployment use private networking? Available through enterprise implementation

Yes, as an option of the deployment kit. In every Azure shape PostgreSQL has no public endpoint and the API has internal ingress only, with one HTTPS entry point at the web tier. The private shape of the Customer Environment kit goes further: it deploys into a virtual network you already run, puts PostgreSQL, Key Vault and the container registry behind private endpoints, gives the application no public address at all (internal ingress), encrypts the database with a key you hold and sends logs to your own workspace. In the standard shape the registry and, by default, Key Vault accept connections from the internet with Entra authentication. An egress firewall is yours to add in either shape; the application needs outbound access to your identity provider's key endpoint and to any signal sources you leave on. The templates compile and lint, and the private shape is set up in your subscription during the implementation.

How are upgrades and rollbacks handled? Available through enterprise implementation

By design: database migrations run before the new code and are written to be backward compatible, so the previous release keeps serving on the newer schema during a rollout, and a rollback restores the previous images without touching the database. Every migration is classified in a registry, and a check fails the build if one is not. The upgrade script lists the migrations an upgrade crosses and stops for confirmation when one is not backward compatible; so far that is three of thirty-nine (0006, 0013 and 0034), each needing a maintenance window, and 0035 refuses going back below release 2.13 while any store limit exists. On Azure a new revision takes traffic only after its readiness check passes. On the single machine an update restarts the containers, typically under a minute. In your own environment you decide when to apply a release. Kelynto Managed is deployed this way on every merged change, with automatic rollback of the images.

Who operates Kelynto in each deployment model? Available through enterprise implementation

There is a shared responsibility matrix for both models. Managed: Kelynto operates the application, database, backups, monitoring, patching and incident response; you run your identity provider, assign roles, send the files and publish the reviews. Customer Environment: you own the subscription, network, keys, backups, monitoring recipients and outbound connections, and you decide when to apply releases; Kelynto publishes releases, fixes application vulnerabilities and supplies the runbook. Day-to-day operation (failed runs, capacity) is shared and agreed in the contract, including who is on call. In your environment your own team is the first responder.

Architecture

Is Kelynto accessible, and which browsers does it support? Not available

There is no WCAG conformance statement or VPAT yet. What is in the code: visible keyboard focus, a skip link, reduced-motion support, labelled controls, and light and dark themes, and everything on the Watch map is also in a list that a keyboard or a screen reader goes through. It runs in the browser with nothing to install on user machines, and it is tested in Chromium, the engine behind Chrome and Edge. The review screen has a narrow layout for phones and tablets, and the app loads no third-party scripts or fonts. If accessibility conformance or a specific browser is a procurement requirement, say so early so it is assessed first against your standard build.

What availability and redundancy does Kelynto offer? Available through enterprise implementation

Availability is built into the production design, and any commitment is made in a contract; there is no standing uptime commitment today. The internal objective is that 99.5 percent of API requests succeed over 30 days. The Azure production settings run two replicas of each application container across availability zones and a synchronous standby database in another zone. The single virtual machine option is the simpler, single-host shape. Kelynto Managed runs on Microsoft Azure in Central US, deployed by an automated pipeline with health checks and automatic rollback. If the service were down on a Monday, the weekly review is produced by a scheduler and an operator reruns it after recovery, and during a pilot your current process stays in place as the fallback.

What are the hosting options, Managed or Customer Environment? Available through enterprise implementation

One codebase, two models. Kelynto Managed: a multi-tenant service on Microsoft Azure, hosted and operated by Kelynto, each customer in an isolated workspace. It runs on Microsoft Azure in Central US, deployed by an automated pipeline with health checks and automatic rollback, and a workspace is opened with each agreement. Customer Environment: the same software runs single-tenant inside your own cloud account, signing in with your directory; you own the account, network, keys and backups. It is available through enterprise implementation, with the Azure deployment kit that runs Kelynto Managed. For AWS there is a Terraform module, and a Docker Compose path for any container host runs the full stack in production mode. Either model is set up as part of an implementation, and the first deployment in your account is validated step by step with your team.

Is the Kelynto service live today? Available today

Yes. Kelynto Managed runs on Microsoft Azure in Central US, in the shape the product's own Azure deployment kit stands up, deployed by an automated pipeline with health checks and automatic rollback of the application images, and its status page is public. A workspace is opened with each agreement: a pilot, a partner sandbox, or the workspaces of a partner's clients, and the first customer workspaces open with the first agreements. Proven by execution: both container images build, the full stack runs in production mode, a complete customer onboarding passes end to end with a test identity provider, and a backup and restore works. Sign-in against a customer's own identity provider is set up and checked step by step at onboarding. To see the product working today, book a 20-minute walkthrough with the founder.

What logging and monitoring is in place? Available through enterprise implementation

In the product: structured JSON logs with a request id on every line, health and readiness endpoints, and Prometheus metrics for request counts, latency, unhandled errors, rate-limit refusals and background runs. Logs carry time, event, request id, workspace and route, sometimes a user's email or id, and never request bodies or sales figures. In the Azure templates: an outside-in availability test and email alerts for database down, server errors, restarts and failed or missing weekly runs, with logs kept 90 days in the production settings. Alerts go to a mailbox; there is no staffed round-the-clock operations team. On a single virtual machine, alerting comes from an external uptime monitor added to the host. In your own cloud account, the alert recipients are yours to name.

What is the architecture of Kelynto? Available today

Four parts. A React and TypeScript web app served by nginx. A Python API (FastAPI) that validates uploads, serves reviews and enforces sign-in, roles and workspace isolation on every request. A scheduler that produces each workspace's weekly review, closes stale runs and purges data past retention. And PostgreSQL, which is the only state. The weekly flow is: validate and load files, pull outside signals, run variance and attribution, build one facts document, write the commentary from those facts, then capture the planner's edits and verdicts. The API keeps no business state between requests, so it can be scaled out. The architecture document and the data-flow document with diagrams are available on request. Kelynto Managed runs on Microsoft Azure in Central US, deployed by an automated pipeline with health checks and automatic rollback.

Where are Kelynto's scale limits? Available today

The local tests cover up to 2,000 stores and 60 categories, 50 concurrent users, and a check with two and three large workspaces in one database, on synthetic data. Shapes beyond that are sized on your own data volume: more stores or categories, more concurrent users, several API replicas, a database on its own host, production hardware on Azure, weekly runs with live signal feeds or the AI writer, and the retention purge on a large workspace. Two slow points are known from the local tests: store-level variance views at 2,000 stores take about 0.2 seconds alone and 0.7 seconds under load, and refreshing per-store signal sources for 2,000 stores means thousands of sequential requests, so it runs as a separate job ahead of the review. A chain much above 500 stores starts with a sizing check on its own data volume.

What scale has Kelynto been tested at? Available today

The measurements are local tests on synthetic data, on a small shared machine: 2 virtual cores, 8 GB, with the database, API and load generator on the same host, each test run once. At 500 stores and 60 categories the weekly run took about 2 seconds; at 2,000 stores about 4.5 seconds, with the template writer and no live signal refresh. Ingest ran at roughly 27,000 to 33,000 rows a second. With 50 simulated users pacing one request about every 5 seconds against a 2,000-store workspace in production sign-in mode, the median response was about 45 milliseconds and the 95th percentile about 0.4 seconds, with no errors. Treat every figure as plus or minus 10 to 15 percent. A chain much above 500 stores starts with a sizing check on its own data volume.

Security

What is sent to the AI model when it is enabled? Available today

For the AI writer, the week's evidence: store, region and category names, the week's amounts and shares, the titles and dates of records on file, and the results of calculations the model asked for. For the in-app assistant's language-model mode, if your admin switches it on: the question asked and the name of the page it was asked from, and no record. The model receives no raw row, person, file or email address, and nothing of another workspace. So the AI writer sees business-sensitive aggregates, including store names, which is why both are off unless your admin turns them on. If that is not acceptable, leave the defaults, which send nothing anywhere, forbid language models for the whole deployment with one setting, or use the Azure OpenAI option with a resource in a subscription you control. Each provider's retention and training terms are checked and recorded before that provider is enabled for a customer.

Does Kelynto use AI on our data? Available today

The engine that explains the week is rule-based, not AI, and it computes every number. The weekly commentary is written by Kelynto's built-in writer, and the assistant inside the product answers by fixed rules. A language model writes the commentary only when the AI writer is switched on for your workspace: by your workspace's admin, or, for a partner's clients, under the partner's recorded agreement with them. A deployment setting can forbid it for every workspace. The AI writer uses Anthropic by default or your company's own model, each a subprocessor only for workspaces that enable it. Every claim it writes is checked against the evidence it cites, and a planner reads and publishes the review; the assistant's model only picks the kind of question, and no wording it writes is shown. It makes no decisions about individuals. Provider terms are checked before you enable one.

Is there an audit log? Available today

Yes. Every state change writes an audit event with actor, action, entity, request id, time and details: uploads, runs, commentary saves, publishing, labels, settings changes with before and after values, user changes, operator commands, scheduled runs and retention purges. A bounded set of reads is recorded too: exports and downloads, admin screens and sign-ins. Admins read, filter, verify and download it in the app. On PostgreSQL the application's database role can only add events, and each event is linked into a hash chain per workspace, so a later change can be detected. That makes it tamper-evident rather than tamper-proof: the database role that owns the tables could rewrite events together with their hashes, which shows against a copy exported earlier. Operator commands are recorded under a service user. It is kept seven years by default.

How do users sign in? Available today

Through your own identity provider using OpenID Connect: the authorization code flow with PKCE, a public client with no secret. The API verifies the signed access token on every request (signature, issuer, audience, expiry), accepts only RS256/384/512 and ES256/384, and decides the workspace from the token's verified issuer, never from a header. Setup is documented for Microsoft Entra ID, Okta and Auth0, and any other standards-compliant OIDC provider is set up the same way. Sign-in is tested end to end with a test identity provider and signed tokens, and your first real token is checked step by step at onboarding. A workspace admin can set single sign-on up in the Admin pages, with a test sign-in before it is switched on, or a Kelynto operator can register your provider. For teams without an identity provider the product also has its own sign-in by emailed link, which the operator switches on.

Can the AI writer use Azure OpenAI in our own tenant? Available through enterprise implementation

Yes. Microsoft Azure OpenAI is one of the models a workspace admin can choose as the company's own for the AI writer, beside Anthropic, OpenAI, xAI Grok, Google Gemini and any OpenAI-compatible endpoint. The admin enters the endpoint, the model deployment and the company's key in the AI writer settings; the key is sealed with the deployment's key and shown only by its last four characters. When it is chosen, the week's evidence is sent to that resource, so inference stays inside the Azure subscription it belongs to. In the Customer Environment model that can be your own subscription and your own resource. By default the AI writer uses Anthropic Claude.

How are backups and disaster recovery handled? Available through enterprise implementation

Backups are part of every deployment shape. In the Azure templates the database has automated backups with point-in-time restore, kept 35 days in the production settings, with an optional copy in a second region and an optional standby in another zone. The single virtual machine option takes a nightly encrypted dump kept 35 days. Recovery figures in the documents are design targets: for example a recovery point of zero and recovery time of 15 minutes for a zone failure with the standby database, and up to 24 hours of data loss on the single machine. What has been tested: a backup and restore of the single-machine procedure, in a local rehearsal with synthetic data. Regional recovery is documented as an outline.

Does Kelynto have SOC 2 or other certifications? Not available

No. Kelynto has no SOC 2 report, ISO 27001 certificate or other third-party attestation, and it has not been assessed against HIPAA, PCI DSS or FedRAMP; its data contract has no field for cardholder or health data. No date is given here for any of these. What is available: the security overview, the Trust page with the status of every control, the review pack with a pre-filled questionnaire, the architecture and data-flow documents, and the option to run the product inside your own cloud account so your own controls apply. The infrastructure the templates target is Microsoft Azure, whose physical and platform controls are covered by Microsoft's own attestations. If an attestation is a hard gate for you, say so in the first conversation.

How is our data deleted when we leave? Available today

A workspace admin can do it from the Admin pages. They make a full export if they want a copy, then schedule the deletion by typing the workspace's name. Nothing is removed for a grace period, seven days by default, and any admin can cancel. Then the workspace is closed and removed in a single transaction that checks nothing with that workspace's id remains in any table, and a deletion certificate is written: what was deleted table by table, when, who asked, and what remains and for how long. That certificate is Kelynto's own record, not a third-party attestation. What remains, and it is stated before you confirm: database backups hold the deleted rows until backup retention passes, and logs hold request metadata, never sales figures, until log retention passes. In your own environment you can delete the resource group.

How is our data protected? Available today

The short version. Sign-in goes through your own identity provider, so passwords and multi-factor authentication stay with you; Kelynto stores no passwords. Each customer's data is separated twice, in the application and again in the database with row-level security that shows a connection carrying no workspace nothing at all. Four roles are checked on every request, and every change is written to an audit log. The product holds weekly sales, plan and in-stock totals by store and category, never shopper, payment or employee data. The AI writer is switched on per workspace, and the engine computes every number either way. Encryption in transit and at rest comes from the platform it is deployed on. Kelynto Managed runs on Microsoft Azure in Central US, deployed by an automated pipeline with health checks and automatic rollback. The Trust page lists every control with its status.

Where is our data stored, and can you guarantee data residency? Not available

Data is stored in one Azure region chosen at deployment, and the region for a customer is stated in the contract; there is no data residency guarantee beyond that today. The templates default to East US 2, and geo-redundant backup, if enabled, adds the paired region. In the Customer Environment model the data sits in your own subscription, in the region you choose. Two things cross the boundary whatever the region: store coordinates sent to weather and event services, and the weekly facts if you enable an external AI writer. Kelynto operates no data centre of its own and runs on Microsoft Azure's regions.

Is our data encrypted? Available through enterprise implementation

Encryption comes from the platform the application is deployed on. In transit, the app is served over HTTPS only with HSTS, and in the deployment kits PostgreSQL refuses connections without TLS. At rest, the kits use the cloud service's encryption for the database, its backups, secrets and logs, with keys the cloud provider manages by default. In a Customer Environment the kit can instead use a key you hold: on Azure for the database and its backups, on AWS for the database, secrets, logs and registry (that module is checked statically). Field-level encryption and a separate key per workspace are not part of the product. On a single virtual machine, the containers talk over the host's internal network, and backups are additionally encrypted to a key kept off the machine.

What can the in-app assistant read, and who sees what I ask? Available today

It reads only the workspace you are signed in to, through the same services and permission checks as the screens, inside the same database isolation: it cannot reach another workspace, and a viewer is told only what a viewer may read. No API key opens it. Conversations are stored per person in your workspace; only the person who asked can open them, and they are covered by your retention period, the full export and the deletion of the workspace. The audit log records who asked and which records were read, not the text of answers. A rating and note you leave on an answer can be read by your workspace's admins and by the operator. By default no language model is involved. An admin can switch the assistant off for the workspace, and a deployment can remove it.

What is your incident response process? Available today

A written plan exists: severity levels, containment steps for suspected data exposure (disable the identity provider, deactivate the workspace or account, take the application offline if needed), evidence preservation, customer notices and a written review afterwards. The notification target is within 24 hours of confirming that your data or service is affected; the contract would govern. The plan has not yet been exercised in a live incident. Alerts go to a mailbox, and there is no round-the-clock team. Security reports go to security@kelynto.com, and the named incident contact for your agreement is set by the founder.

What are the known limits of the tenant isolation? Available today

They are listed openly. The workspace a transaction is bound to is an ordinary session setting, so a flaw that let an attacker run arbitrary SQL as the application could set it to each workspace in turn. The application connects as a database role that owns nothing, so the same flaw could not switch row-level security off or alter a policy; the separate owner role, used for migrations, can, and access to it rests on who holds its credential. The table of workspaces (names and settings, no sales data) is readable by the application role. Record ids are sequential across workspaces, so a customer could estimate overall volume. Operators with database or container access can see all workspaces. No independent party has verified the isolation yet. These are in the security overview under Not yet in place.

What security controls are not in place yet? Available today

The security overview keeps an open list, and the trust page labels every control. The main items: no third-party attestation and no external penetration test. Self-service single sign-on and SCIM are built and tested with stand-ins for Entra ID and Okta, and a real tenant is connected for the first time at a customer's onboarding. Tokens are not checked for revocation with your provider. The emailed sign-in's second factor is an authenticator app. Roles are the four built-in ones. The workspace binding in the database is a session setting, and the owner database role is not limited. Rate limits are per process. Background runs have no durable queue. Uploaded files are parsed as data and discarded, with no virus scan. A bounded set of reads is audited, and the audit log is tamper-evident. No field-level encryption, no egress firewall in the kit, no push to a SIEM. The retention purge leaves backups to expire. Browser testing is in Chromium.

Who at Kelynto can see our data? Available through enterprise implementation

Technically, in the managed model, anyone with operator access to the production environment could: opening a shell in the API container gives access to every workspace's data, and shell sessions are not recorded by Kelynto. Control-plane actions are in the Azure activity log, and operator commands that change data are written to your audit log, under a service user, not a named person. The names and number of people with access, staff screening and whether administrative access requires multi-factor authentication are organisational facts the founder states in the review pack. If this is a concern, the Customer Environment model changes the answer: Kelynto staff see your data only if you grant access to your subscription.

Can I see another customer's results? Not available

No. Each customer's workspace is isolated in the application and in the database: a request for another workspace's record returns not found, and one customer's data, labels and Score and learn figures are never shared with another. The assistant on this site has no access to any workspace data at all. What you can see is the demonstration workspace, Harborview Markets, which is synthetic and labelled as such, and a backtest on your own past weeks.

What security documentation can you share? Available today

Four things exist today. A security overview written for a retailer's security and IT review, covering data handled, sign-in, workspace isolation, audit, AI use, platform controls and a plain list headed Not yet in place. A control register, shown on the Trust page of this site: every control with one status (implemented, available through enterprise implementation, planned, not in place) and the file or test that shows it. A security review pack: a pre-filled questionnaire in the style of CAIQ Lite and SIG Lite, architecture and data flows with diagrams, a subprocessor template, an incident response and recovery plan, and a shared responsibility matrix. And the deployment document with a record of what was proven by execution. The founder completes the organisational answers in the pack before it is sent. Ask for the pack and name the reviewer.

Does Kelynto store passwords, and is MFA enforced? Available today

No passwords are stored, in any sign-in path. With your own identity provider, the application only verifies signed tokens, so multi-factor authentication and conditional access are enforced by your provider under your policies, and Kelynto inherits them. The product's own sign-in, for teams without an identity provider, uses an emailed single-use link or code, stored only as hashes. A person can add an authenticator app as a second factor (time-based codes, with ten single-use recovery codes shown once and stored hashed), and a workspace admin can require it for every member. Five wrong codes lock the second step for fifteen minutes, and each event is in the audit log. The built-in sign-in is off in production until switched on. There is no emergency login that bypasses these; an operator can remove a lost authenticator from the command line, which is audited. A development header sign-in exists for local demos, and the service refuses to start with it in production mode.

Has Kelynto had a penetration test? Not available

No. There has been no third-party penetration test and no external security audit. What has been done is internal: an independent QA and security pass inside the project, run against the release, covering cross-workspace access, role escalation, token attacks, upload abuse, injection and server-side request forgery, with the open findings fixed or listed. It is an internal pass, not an external test. Whether a customer runs its own test, and against which environment, is agreed in the contract. Automated dependency and image scanning runs in the build pipeline on GitHub, since October 2026.

Is there rate limiting and abuse protection? Available today

Yes, in the application: 600 requests a minute per signed-in user, 60 a minute per client address for requests without valid credentials, and 20 a minute for uploads and run triggers, each with a burst allowance. A refused request gets a 429 with a retry time. An upload's token, role and limits are checked before its body is read, and the size limit is applied as it arrives. The counters are per API process, so they are the inner layer. A web application firewall is available in the Azure templates through Front Door, off by default, and the AWS module puts AWS WAF in front of its load balancer; there is no dedicated DDoS protection plan in the templates. The early upload refusal is verified behind the supplied nginx.

How long does Kelynto keep our data? Available today

Retention is set per workspace by an admin, with a separate length for each kind of data: weekly figures, weekly reviews, outside signals, loaded file records and the audit log. Business data is kept 730 days by default (minimum 30). A daily purge deletes what is older and writes an audit event with the counts. Before a change is saved, the admin sees what the next purge would remove, and after a length is shortened the purge waits a day. Stores, user accounts, identity provider settings and workspace settings are kept. The audit log is kept 2,555 days by default (seven years, minimum 365) and never less than the longest of the others. The purge works on the live database; backups keep rows until the backup retention has passed, because backups cannot be edited selectively.

How are roles enforced and how fast is access revoked? Available today

Four roles (viewer, planner, analyst, admin) are checked on every API route, not just hidden in the interface. With your own identity provider, roles come from the token's roles claim: an account is created the first time a user with a role signs in, and removing the role in your directory revokes access on the user's next request. A deactivated account in Admin always wins. Your identity provider can also deactivate members directly through the provisioning service, which closes their sessions and revokes the API keys they made. An admin can see in Admin, Roles exactly what each role and each person can do. Two settings to make: a token already issued stays valid until it expires, so access token lifetimes should be an hour or less; and accounts are matched by email, so the email claim should come from an attribute users cannot edit.

Do you support SAML single sign-on? Not available

No. Single sign-on uses OpenID Connect: the authorization code flow with PKCE, which Microsoft Entra ID and Okta both offer for the same users and groups. The documentation has the setup steps for each, and a workspace admin can connect the provider in the Admin pages and run a test sign-in before switching it on. No date is given for SAML.

Can our identity provider provision and deactivate users? Available today

A provisioning service that follows the SCIM 2.0 standard is built. A workspace admin makes a provisioning token in the Admin pages and enters it, with the address shown, at the identity provider. The provider can then create members, keep names and addresses current, and deactivate them. A member deactivated this way is signed out at once and the API keys they made stop working. Directory groups give roles where the admin maps a group's name to a role, and the screen shows who would change role before the mapping is saved. The workspace owner and the last admin cannot be removed this way. It is tested with the requests Microsoft Entra ID and Okta document, and the first connection to your provider is checked with a test group.

Do you offer an SLA or support hours? Not available

No SLA, service credits or defined support hours today; any commitment is made in a contract. Kelynto is founder-led: during a pilot the same person is the pilot lead and, in the managed model, the operator, and alerts go to a mailbox. There are internal objectives, for example that 99.5 percent of API requests succeed over 30 days, and the recovery point and time figures are design targets. A pilot includes a kickoff, reading the validation reports with your data contact, a weekly check-in and a readout. If a commitment is a precondition, raise it in the first conversation.

Can we set up single sign-on ourselves? Available today

Yes, in the Admin pages. An admin enters the issuer address, client ID and audience of an OpenID Connect provider. The product reads the provider's published configuration and says in words what is wrong if it cannot use it. Nothing is switched on until a test sign-in passes: the admin signs in at the provider and the token is checked exactly as every later request will be. To send people who type their address to your provider, the admin proves control of the email domain with a DNS record. Single sign-on can then be required, which refuses email codes for members. One named admin keeps email sign-in so the workspace can still be reached if the provider is down, and each such sign-in is recorded. It is tested with stand-ins for the provider and for DNS, and the test sign-in checks your own provider before anything changes.

Who are your subprocessors? Available through enterprise implementation

The subprocessor register is a template per hosting option, and the list for your deployment, with regions and how changes are notified, is fixed in the contract. What each option implies: in the Kelynto Managed model on Azure, Microsoft holds everything stored. A language model provider (Anthropic by default, or the provider of your company's own model) and PredictHQ apply only if your admin switches those features on. A mail delivery provider applies only where sign-in by emailed link is switched on. Customers on a shared sign-in provider add that identity provider. Weather services receive store coordinates only and are listed for transparency. In the Customer Environment model the software runs in your account, Kelynto holds none of your data and engages no subprocessor for it.

How is our data kept separate from other customers? Available today

In the managed model customers share one application and one PostgreSQL database, separated by two independent layers. In the application, every query from a workspace-bound session is filtered by workspace and writes for another workspace raise an error. In the database, row-level security is forced on all eighty-one customer-owned tables for reads and writes, and the application connects as a role that cannot bypass it and does not own the tables, so it cannot switch it off. The policy fails closed: a transaction bound to no workspace sees no customer rows and can write none. A request for another workspace's record returns not found. Automated tests cover both layers, including raw SQL. If shared infrastructure is not acceptable, the Customer Environment model is single tenant in your own account.

Do you train AI models on our data? Available today

No. Kelynto does not use customer data to train any model. Attribution is a rules engine, so there is no model of ours to train. The labels your planners create are used inside your own workspace only, stay in your workspace and can be exported by you. If your admin enables an external AI writer, the provider you choose receives the week's aggregated facts, and that provider's own API terms govern what it does with them; those terms are checked and recorded for each provider before you enable it. Any use of anonymised learnings across customers would have to be agreed in your contract; nothing in the product does it.

How do you manage vulnerabilities and secure development? Planned

Built, configured and running on GitHub since October 2026. Python packages are pinned to exact versions used by both the tests and the image; JavaScript packages are locked; build actions are pinned to commit hashes. The build workflows audit dependencies, scan both container images and the git history for secrets, block deployment on a critical vulnerability with a fix available, and attach a software bill of materials. Run by hand on 2 October 2026, the dependency audits found no known vulnerabilities. The status label is planned because the operating process around the tooling comes next: recorded scan results for the release images, a written patching deadline by severity, and base images pinned by digest. An automated backend test suite of more than 800 tests at version 2.3.0, including database isolation tests, has to pass before a release.

Commercial

How do I talk to a person at Kelynto? Available today

Two ways, both reach the founder directly. Book a 20-minute conversation at https://calendly.com/kelynto/30min, or write to marc@kelynto.com. Marc Estinville, the founder, answers personally. For a first conversation it helps to say your role, roughly how many locations you have, what your plan is built in, and what prompted the question. For a security or procurement review, say who the reviewer is and which hosting model you are considering, so the right documents are sent. Security reports go to security@kelynto.com.

What contract and legal documents do you have? Planned

Drafts are ready for review by counsel: a mutual NDA, a supervised pilot agreement, a data processing addendum, a customer-hosted access statement and a pilot success criteria form. The founder finalises them with you, and this guide does not quote their terms. The onboarding checklist expects the pilot agreement to name retention periods, deletion on exit and incident contacts on both sides, and the shared responsibility document lists what a customer should have written in: hosting option and region, subprocessors, retention and deletion, incident notification, any availability commitment, and whether an AI model may be enabled. There is no master services agreement or published terms of service yet; contract terms beyond the pilot are agreed with the founder. Questions about your paper or your NDA go to the founder. Nothing here is legal advice.

Can we create our own workspace and load data ourselves? Planned

The self-service route is built, and it opens with the hosted service. In the product, a person signs in with an emailed link or code, creates a workspace for their organisation (name, time zone, first day of the fiscal week, currency), becomes its first admin, invites colleagues by email with a role, and loads data through an upload preview that shows how each column is read before anything is stored. Whether sign-up is open, by approval or by invitation is set by whoever operates the deployment. Today a Kelynto operator creates the workspace as part of onboarding, and the way to start is a pilot conversation. This answer changes when the site reports sign-up as open.

Do you have a data processing agreement? Not available

A draft data processing addendum is ready for review by counsel, and whether Kelynto signs a customer's own instead is a founder decision. The facts such an agreement needs are documented: the categories of data (weekly business aggregates and users' business contact details), the subprocessor template per hosting option, retention and deletion behaviour, and the incident notification target. In the documents Kelynto's role is processor and the customer is controller. Standard contractual clauses and cross-border transfer terms are drafted with the founder when a deployment needs them. If a signed agreement is a precondition, raise it at the start so it is settled before the first upload. This is not legal advice.

How much does Kelynto cost? Planned

Pricing is not finalised. It will depend on deployment model, number of locations and pilot scope. The fastest way to get a number for your situation is a 20-minute conversation.

Terms are set for your situation rather than read from a price list, so the assistant does not estimate, hint at a range or negotiate. A number that did not come from the founder in writing is not a Kelynto price. To get terms for your situation, book 20 minutes at https://calendly.com/kelynto/30min or write to marc@kelynto.com, and bring your number of locations and preferred deployment model.

What is on the roadmap? Planned

Kelynto does not publish a dated roadmap, and the assistant gives no delivery dates. What it can say precisely is where things stand. Built and tested, ready for the hosted service: the one-click sample workspace, self-service workspace creation with sign-in by emailed link, and invitations. Built for deployment in your cloud: the Azure deployment kit, which runs Kelynto Managed, and the AWS Terraform module. Built for partners: the partner API, embedded views in the partner's brand, and client licences as a pool. The security overview keeps a list of next security items, such as passkeys for emailed sign-in and shared rate-limit counters; it is a record of work, with no dates attached. The changelog shows what has shipped, release by release. If a specific capability decides things for you, tell the founder.

Is there a sample workspace I can explore? Planned

It is built, and it opens to visitors with the hosted service. The sample workspace is a private copy of Harborview Markets, a synthetic retailer with 36 stores: a story in each of the latest weeks, draft reviews you can edit and label, Watch with its map and list, and a Score and learn page, which counts real use only and so stays empty on synthetic data. A visitor opens it from the site in one click, without an account, can look through each role, and the copy is deleted after 24 hours. The site shows the link as soon as the product reports the sample workspace as on. Today, the product tour on the site shows the product's screens with the same synthetic retailer, and the founder can walk you through it live in 20 minutes. It is labelled as synthetic everywhere: its numbers show the experience, not results.

Can I try Kelynto without talking to sales? Planned

Two self-service routes are built and open with the hosted service: a one-click sample workspace filled with synthetic data, and creating your own workspace with sign-in by emailed link. The site shows both links as soon as the product reports them as on. Today, without a call, you can take the product tour on the site, which shows real screens of the product with the synthetic demonstration retailer, read the deployment and trust pages, and ask this assistant. To see it working on your own questions, book a 20-minute conversation at https://calendly.com/kelynto/30min. You talk with the founder directly, with no sales process in between.

Company

Who is behind Kelynto? Available today

Kelynto was founded by Marc Estinville, who has more than 17 years in retail supply chain and planning technology. The product comes from that work: the weekly task of explaining to leadership why sales missed or beat plan, which often falls to a planner and a spreadsheet. In a pilot you work with Marc directly. His background is context for why the product is shaped the way it is. You can reach him at marc@kelynto.com or book 20 minutes at https://calendly.com/kelynto/30min.

What does the name Kelynto mean? Available today

Kelynto is a coined name. That is all we say about the name itself. It is written as one word with a capital K, and the website is https://kelynto.com. What the name refers to is easier to describe: Kelynto is variance commentary for retail planning, the weekly review written from your own records, and its tagline is "Your Monday business review, already written."

What happens to our data if Kelynto shuts down? Available through enterprise implementation

A fair question for a founder-led company. Your raw sales and plan data come from your own systems, so you always hold them. What Kelynto adds, the commentary, cause labels and Watch outcomes, can be exported by your analysts and admins at any time, and a full export of the workspace is one archive. Your planning process keeps running on its own, so if Kelynto stopped, you would return to how you work today. In the Customer Environment model the running software and all data sit in your own cloud account. Source code escrow, a third-party continuity arrangement and a wind-down period are contract terms: there is no standard arrangement for them yet, and they are agreed with the founder. Raise it with the founder and expect a straight answer.

How established is Kelynto as a company? Available today

Kelynto is founder-led, with a complete working product: the whole weekly loop, the partner API and embedded views, and a Trust center that lists every control with its evidence. Kelynto Managed runs on Microsoft Azure in Central US, deployed by an automated pipeline with health checks and automatic rollback. The first customer workspaces open with the first agreements, and no customer is in production yet. What that offers you is direct access to the person building it, proof on your own history before you buy (a pilot starts with eight sealed weeks, judged on measures agreed up front), and the option to run the software in your own Azure account. Details such as team size and funding are answered by the founder, in writing if you need them for a vendor file.

How big is the team and how is Kelynto funded? Not available

Kelynto is founder-led. Headcount, investors, funding, runway and office location are answered by the founder directly, in writing if you need it for a vendor file; the assistant states only verified facts and leaves these to him. Those are reasonable diligence questions. Write to marc@kelynto.com or book 20 minutes at https://calendly.com/kelynto/30min.

How can we check Kelynto before we rely on it? Available today

The product is set up so you can check it rather than take it on trust. Your current process stays in place: Kelynto sits above your planning system and writes nothing back. A backtest on your own past weeks shows the fit before anyone changes how they work. The pilot is judged on measures counted from your planners' own actions, on the Score and learn page. The security overview and the trust page give the status of every control. You can export your labels and commentary at any time, and deletion on exit is a documented procedure. If you host it in your own cloud, the software and data sit in your account. What it asks of you is your team's attention for a few weeks.

What is this assistant, and what happens to what I type? Available today

This is an automated assistant, not a person. It answers questions about Kelynto from an approved knowledge base, each answer with a status label, and it has no access to any workspace or customer data. It works in one of two ways. In browser mode, as on this site today, your question is matched in your browser against answers published with the site: nothing is sent or stored. When it is connected to the product's service, conversations are stored so the team can see which questions go unanswered. As built: the client address is never stored (only a keyed hash that changes every day, for abuse control), no cookies are read or set, and email addresses, phone numbers, card-like numbers and credentials are removed before storage. By default conversations are kept 180 days and contact details you submit with consent 730 days. To reach a person or have your data removed, write to marc@kelynto.com.

Who else uses Kelynto? Not available

Nobody yet. The first customer workspaces open with the first agreements, so no retailer is named and there are no case studies or reference customers to point to. In place of a reference you can check it yourself: the product tour on a synthetic retailer, a backtest on eight sealed weeks of your own history, set beside what your team actually wrote, and a pilot judged on three measures agreed before it starts. A partner starts with a sandbox of synthetic clients. If a reference customer is a requirement for you, say so in the first conversation.

Not answered here? A person will answer.

Email marc@kelynto.com Book a 20-minute conversation