AI Readiness Scorecard

Find out where
you actually are.

Fifteen questions across five dimensions, about five minutes, and a result you could put in front of a board without translating it first.

Your answers stay in this browser. Nothing is sent anywhere, and you get the result before anyone asks you for an email address.

Begin↓

What it measures

01
Direction
Whether anyone senior can state what data and AI are actually for in your business.
02
Accountability
Whether ownership is written down, or shared in the way that means nobody has it.
03
Foundations
Whether the data underneath would survive being pointed at a model.
04
Delivery
Whether work reaches production and gets used, or reaches a demo and stops.
05
Capability
Whether the people you have match the outcomes you need, titles aside.

01Direction

Whether anyone senior can state what data and AI are actually for in your business.

1If I asked three of your executives what data and AI are for in this business, what would I hear?

Three different answers, and at least one would be a tool nameBroadly the same theme, but nobody could say what it changesA consistent answer tied to a business goalThe same answer, with the numbers it is expected to move

2How is the data and AI agenda decided?

Whoever asks loudest, or whoever is closest to the teamAn annual plan that reality overtakes by about MarchA prioritisation forum that mostly holdsA named executive owns the sequence and defends it

3What happens when someone asks for something not on the plan?

It gets done, because saying no is harderIt joins a backlog nobody revisitsIt is triaged, and sometimes displaces somethingIt goes through demand management, and the trade-off is explicit

02Accountability

Whether ownership is written down, or shared in the way that means nobody has it.

4Who owns data quality in your business?

Everyone, which in practice means nobodyThe data team, whether or not they can control itThe business owns it, with the data team supportingNamed owners per domain, written down and reviewed

5Is there anything in writing between the data function and the rest of the business?

Nothing. It runs on goodwillA service catalogue or a ticket queueInformal agreements that mostly holdA signed agreement per domain, with commitments on both sides

6When two reports disagree in front of the board, what happens next?

An argument, and the meeting moves on unresolvedSomeone goes away to reconcile it, and it recursThere is an agreed source for the important numbersIt rarely happens, because certified definitions exist and are used

03Foundations

Whether the data underneath would survive being pointed at a model.

7How long would it take to answer a new question that spans two parts of the business?

Weeks, and the answer would be contestedDays, with a specialist doing it by handA day or two, using existing modelsHours, and someone non-technical could do it

8Could someone find out what a given data set means and whether it can be trusted?

Only by asking the person who built itDocumentation exists somewhere and is out of dateA catalogue exists and covers the important thingsDefinitions, ownership and quality are visible where the data is used

9If you pointed a model at your customer data tomorrow, what would happen?

It would learn from duplicates and gaps, confidentlyIt would work after a lot of manual cleaningIt would work for the well-governed subsetIt would work, and we would be able to explain what it did

04Delivery

Whether work reaches production and gets used, or reaches a demo and stops.

10What proportion of data and AI work reaches production and stays there?

Very little. There are a lot of pilotsSome, and it needs babysitting afterwardsMost of what we startMost, and we retire what stops being used

11How do you know whether something you shipped is being used?

We do not measure thatAnecdotally, if someone complainsUsage is tracked for the main productsAdoption is a stated target with an owner

12How do the data people and the business people work together?

Requests in, output out. Little contactProject by project, re-formed each timeStanding cross-functional squads for the big areasSquads with defined roles and a written way of working

05Capability

Whether the people you have match the outcomes you need, titles aside.

13Do your job titles describe the work people actually do?

Almost everyone is an analyst or a scientist regardlessRoughly, though the boundaries are blurredMostly, with real distinctions between rolesYes, and each role has a written definition of what good looks like

14If your most knowledgeable person left tomorrow, what would break?

A great deal. Much of it is in their headSome things, and we would find out which the hard wayLittle. The important things are documentedLittle, and we have deliberately built the cover

15Can people outside the data team answer their own questions?

No. Everything comes back to the teamA few power users canMany can, within governed boundariesMost can, and they are trained and supported to

0 of 15 answered

Your result

out of 100, across five dimensions

Direction

0/9

Accountability

0/9

Foundations

0/9

Delivery

0/9

Capability

0/9

What this means

The first move

Want this written up properly?

Leave an email and you will get your result as a short written read, with what the pattern usually means and what tends to work next. One email. Nothing else.

First name optional

Organisation optional

Work email

Your answers are not sent with it. Only your band.

Company website

A number is a start. A conversation is faster.

If the result matched what you already suspected, that is usually the moment to do something about it. Thirty minutes, no deck.

Book a discovery call→See the advisory packages