The AI-Ready Data Standard · Version 1.0 · September 2026

Twenty-five checks between you and an answer worth acting on.

Your AI will answer every question you ask it about your business. That is the problem. It answers the ones it should refuse, in the same confident voice, and nothing on the screen tells you which is which.

These twenty-five checks tell you instead. Each one takes a minute. Tick what you can prove today and watch your score move.

Start scoring
0/ 25 Tick a check to begin. Half-done counts as zero.

A check passes only if you can show it. Not if you believe it.

"We know our costs" is not a pass. Opening the file where the costs live, with dates on them, is a pass. The difference matters because half-done is precisely what produces a confident wrong answer, and a confident wrong answer is worse than no answer at all. You act on it.

So score honestly. One point for each check you could demonstrate to somebody this afternoon. Zero for anything you would have to build, find, or go and ask about first.

Nothing you tick is sent anywhere. The score lives in this browser and nowhere else.

Defined

Checks 1-5 · every number means one thing

Ask three people for last month's revenue and you should get one number. Often you get three, and all three are defensible.

01One written definition per metric

Ten metrics, one sentence and one formula each: revenue, new customers, repeat rate, ad spend, return on ad spend, contribution margin, units, average order value, refunds, inventory on hand.

TestAsk two people in different teams what net revenue means here. Same answer?
If notEvery comparison between two periods is unverifiable.

02Every revenue number states its basis

Orders placed, money settled, or store gross sales. Named on the number itself, not in a footnote nobody opens.

TestTake last month's revenue from any report. Which of the three is it?
If notOne month reads three ways and every reading is defensible.

03A written close rule

The date each source stops changing. For example: a month is closed ten days after it ends.

TestIs yesterday's revenue final? On what date does last month become final?
If notAn unfinished month gets compared to a finished one and reads as a collapse.

04Fixed period boundaries

One time zone per channel, written down, including the marketplace clock your sales are stamped in. One currency, with the conversion rule if you sell in more than one.

TestDoes "yesterday" mean the same hours in your store report and your marketplace report?
If notYour best day moves depending on who runs the report.

05Definitions are dated, and changes are visible

A definition that changed in June says so, so a year-on-year comparison can be trusted or discarded on purpose.

TestWhen did the definition of a new customer last change?
If notA metric quietly changes meaning mid-year and the trend is fiction.

Connected

Checks 6-10 · the things are linked

One bottle of shampoo wears three codes: a marketplace product code, a store variant, a barcode on the shelf. Nothing in your data says they are the same bottle, so your best seller arrives as three modest products.

06One product family map

Every sellable code maps to the product family a customer would name. One file, one owner.

TestList every code in your best-selling family. One place, one minute.
If notThe ranking of your own products is wrong, and reads as fact.

07Units per pack recorded

A 3-pack is one order and three units. Both numbers are knowable, on every row.

TestHow many units did you sell last month, and how many packs?
If notReorder plans built on pack counts run you out of stock.

08Orders join to customers, and customers to a first-purchase date

Every order belongs to somebody, and every somebody has a start date, even where the marketplace hides the person behind an alias.

TestHow many customers first bought in June, and what have they spent since?
If notLifetime value becomes a formula instead of a measurement.

09Ad spend joins to what it sold

Each campaign points at a product or family and a destination, so spend can be scored against the sales that actually happened.

TestWhich campaigns send shoppers to the marketplace rather than your store?
If notThe channel that drove the sale gets none of the credit, and gets cut.

10Retail and wholesale codes tie back to the same families

Or you can state plainly which sales are missing from every roll-up, and everyone knows it.

TestDoes "how is this family doing across channels" include store shelves?
If notA whole channel silently disappears from every total.

Labeled

Checks 11-15 · every number carries its context

A number with no label is a trap, and the trap is sprung by whoever trusts it most.

11Every figure states its source

Which system produced it, not which dashboard displayed it.

TestPick any number in your weekly report. Which system recorded it?
If notNobody can check the number, so nobody does.

12Every period is marked closed or in flight

Visibly, wherever it appears, including inside the answer an AI gives you.

TestDoes the current month look different from a closed one at a glance?
If notUnfinished data gets read as a trend and acted on.

13Known blind spots travel with the number

The ad pixel that cannot see marketplace checkouts. The buyer identity that arrives late. The report that leaves a channel out.

TestDoes your ad report say what it cannot see?
If notA profitable channel gets cut because one tool was blind to it.

14Immature cohorts are blank, not estimated

A group of customers too young to have a repeat rate shows nothing. Not a small number. Nothing.

TestWhat is the 90-day value of customers who first bought six weeks ago?
If notA guess in a table of measurements gets read as a measurement.

15Sources that restate are marked as such

Ad platforms keep crediting sales after a week ends. Marketplace settlements arrive later and correct what you already reported.

TestBy how much did last week's ad return move after the week closed?
If notCampaigns get cut on numbers that were always going to improve.

Owned

Checks 16-20 · the facts only you know are written down

The most important numbers in your business are not in any system you can connect. They are in a founder's head, a supplier email, and a spreadsheet somebody renamed.

16Landed unit cost for every active product, with effective dates

What it costs to make it, ship it, clear customs and prepare it, per unit, and from what date.

TestWhat did this product cost you last month, and when did that change?
If notMargin is calculated with last year's cost, or with none at all.

17Facts have an owner and a recorded approval

Costs, launch dates, targets, which ads point where. Each with a person's name and a date attached.

TestWho approved the cost figure your margin report is using?
If notA number typed into a chat months ago quietly drives every decision.

18Metrics are never stored as facts

Lifetime value and return on ad spend are recalculated every time they are asked for. Never saved.

TestIs any ratio in your systems stored rather than computed?
If notA stale ratio outlives the quarter it was true in.

19Targets are yours, and written down

The number you consider healthy, set by you, not an industry average from an article.

TestWhat payback period counts as good here, and who decided?
If notA borrowed benchmark grades a business it was never measured on.

20Someone owns the map, and reviews it

A named person, a review date, and the habit of adding new products the week they launch.

TestWho added your most recent launch to the family map, and when?
If notThe map is right the day it is built and wrong three months later.

Checked

Checks 21-25 · answers are tested before they are used

The first twenty checks make good answers possible. These five are what stop a bad one reaching a decision.

21A source of record per money number

The system that recorded the money, and it ties to the bank and the books.

TestDoes last closed month's revenue tie to what arrived in the bank, inside a tolerance you wrote down?
If notYou are reporting a number nobody ever received.

22Disagreements are flagged, never averaged

With a written threshold, so two tools disagreeing becomes a message instead of a silent compromise.

TestWhen two reports disagree by 4%, what happens?
If notThe answer is a third number nobody recorded.

23Figures are re-derived before an answer ships

The number in the sentence is checked against the system that produced it, before anyone reads the sentence.

TestTake a figure from an AI answer and recalculate it yourself. Same?
If notA plausible-looking figure goes into a board deck.

24Implausible figures are refused, not rendered

A report that cannot compute something says so, rather than printing a zero or a blank that reads as a real result.

TestWhat does your reporting do when a source fails? Blank, zero, or a stated error?
If notA failed read looks exactly like a bad month.

25A scheduled check fires when an assumption stops being true

Not a person remembering. A job that runs on its own and complains.

TestIf a channel stopped syncing on a Tuesday, what would tell you, and when?
If notThe data is wrong for five months and every report looks normal.

What your score means

Score honestly. The list is only useful if the gaps it finds are real ones.

0-9

An AI will answer every question you ask and you cannot tell which answers are wrong.

10-17

Safe for description. What happened, roughly. Not safe for decisions about money.

18-22

Safe for most decisions, with named gaps you can say out loud in the meeting.

23-25

An answer can be acted on the same day it is given.

Copy this standard, adapt it, publish your score. No permission needed, no attribution required, though we would like the link.

Our own score

A checklist whose author passes everything is an advertisement.

So here is ours, honestly. We wrote the list; we do not clear all of it.

15

Shipped by default. Definitions shared by every answer, labels on every number, disagreements flagged rather than averaged, brand-by-brand isolation, and a weekly job that complains when an assumption stops holding.

6

Yours to decide, ours to hold. Your product family map, your costs, your targets, your close rule. We give them one home and apply them everywhere, but you set them. Nobody can outsource what a product costs.

4

Not yet. Some relationships in the model cannot be explored yet. Facts captured automatically from a session skip the approval queue. Our implausible-figure gate covers reports, not file exports. And a wrong answer does not yet leave a permanent automated test behind.

We publish this because the fourth column is the roadmap, and because a standard nobody fails is not a standard. If you find a claim of ours that overstates what we do, tell us and we will change the page.

Score it, then test it.

Pick five questions you already know the answer to, where the answer is surprising. A product that sells well and loses money. A channel that looks bad and is not. Then ask an AI connected to your data. If it only gets the easy ones right, the twenty-five above tell you which piece is missing.

Bring one of those questions to a call