AI Data Readiness Assessment

Your AI pilots work. Your data doesn’t.

In 2 weeks, you’ll know whether the data behind your AI use case is ready for production, what is holding it back, and in what order to fix it.

See if my use case fits

We reply within 2 business days. Two weeks, remote, at a fixed price.

An example readiness profile, built from fictional data. Select a branch to see the finding.

AI is new. The reasons it fails aren’t.

Most AI and agent projects don’t stall because of the model. They stall because the data underneath isn’t ready for it. If any of this sounds familiar, you’re not alone:

  • Your AI assistant answers with last quarter’s numbers.
  • Nobody can tell you which table is the source of truth.
  • Every new use case starts with weeks of data cleanup instead of building.
  • A schema change upstream quietly broke your agent, and you found out from a customer.
  • Your team trusts the pilot demo, but not enough to put it in front of clients.

Gartner predicts that through 2026, organizations will abandon 60% of AI projects that aren’t supported by AI-ready data.

The model is rarely the bottleneck. The foundations are.

Every month they stay unfixed costs engineering time, delays launches and erodes leadership’s trust in AI.

The AI Data Readiness Assessment

Two weeks to know exactly what’s blocking your AI, and what to fix first.

A fixed-price engagement scoped to one AI use case, not your entire data estate. We analyze the data that use case depends on and show you what’s ready, what isn’t, and why.

What you walk away with

A recommendation for leadership
Two pages that answer one question: can we move forward with this use case? Yes, yes with conditions, or not yet, with the three reasons that weigh most.
A scorecard with evidence
The five dimensions scored from 1 to 5. Every score comes with the evidence behind it.
A data map of your use case
The sources, tables and pipelines your use case depends on, with who owns each one and how fresh it is.
Ranked blockers and a 90-day roadmap
Each blocker with its root cause and fix, ordered by impact, and a plan with suggested owners and effort. Also as a CSV ready to import into Jira or Linear.
Data contract recommendations
What to agree on for up to 3 of your most critical datasets: schema, freshness, ownership and quality.
A results session and a 30-day check-in
We walk your team and leadership through the findings. A month later, a 45-minute check-in to review progress before you go live.

Four steps, no surprises.

  1. 5 minutes

    Request

    Tell us about your AI use case in a short form. We reply within 2 business days with next steps, or tell you if the Assessment isn’t the right fit.

  2. Week 1

    Kick-off

    We agree on the AI use case in scope, the systems involved and who we’ll talk to.

  3. Weeks 1–2

    Analysis

    We review quality, ownership, freshness, contracts and how your AI consumes the data, through read-only access and short interviews.

  4. End of week 2

    Delivery

    We present the findings, the ranked blockers and the 90-day roadmap to your team. A month later, we check in on progress.

What we need from you

Read-only access to the relevant systems, and about 6 hours of your team’s time over the two weeks. 2 weeks, remote, English, French or Spanish, NDA signed before kick-off.

Built for teams that are past the demo.

This is for you if

  • You have AI or agent pilots that don’t behave reliably.
  • You want to move to production, but you don’t fully trust your data.
  • You need clarity before investing more in AI.
  • You want a diagnosis with a fixed scope and price, not an open-ended consulting project.

It’s not for you if

  • You don’t have a defined AI use case yet. Start there first.
  • You’re looking for a team to build your AI model. We fix the foundations it runs on.

The Cognitoak method

Five dimensions. One clear answer.

The same five branches from the top of this page. Every Assessment scores them with the same framework, so results are consistent, comparable and backed by evidence.

Quality
Is the data accurate, complete and consistent enough for your use case?
Ownership
Does every critical dataset have someone accountable for it?
Freshness
Does the data arrive on time for what your AI needs to do?
Contracts
Are there explicit agreements between the teams that produce data and the ones that consume it?
Consumption
Can your AI find, understand and access the right data, with the right permissions?

One scale for every dimension

  1. 1Not in placeNo one manages it.
  2. 2Ad hocFixed when something breaks.
  3. 3DefinedClear rules and a named owner.
  4. 4MeasuredTracked and reviewed regularly.
  5. 5Measured and enforcedChecked automatically, with alerts.

At riskReady for production from 3

Built on years of data management in highly regulated industries like banking, where data has to be right.

Your data stays in your hands.

What we commit to in every Assessment, in writing.

Read-only access
We never write to your systems, and the access is removed when we finish.
Under NDA
Signed before kick-off and covering everything we see.
Queried in place
We analyze data where it lives. No extracts or copies on our machines.
Deleted on request
Our working notes are deleted after delivery if you ask.

Questions teams ask before requesting.

How much does it cost?

Each Assessment has one fixed price, agreed in writing before we start. We share it after reviewing your request, because it depends on the systems involved.

What exactly do we receive?

A two-page recommendation for leadership, a scorecard with the evidence behind each score, a data map of your use case, the ranked blockers with a 90-day roadmap (also as a CSV for Jira or Linear), and data contract recommendations. You get the report as a PDF with the editable files, we walk your team through it in a results session, and we check in on progress a month later.

How much of my team’s time does it take?

About 6 hours over two weeks: a kick-off, a few short interviews and the results session.

When do the two weeks start?

Once read-only access is in place. If something on your side moves, such as access or an interview, the dates move with it.

Do you need access to our data?

Read-only access to the systems in scope, under NDA. Your data never leaves your environment.

Does it work with our stack?

The Assessment looks at your data, ownership and processes, so most of it doesn’t depend on your tools. We work with any source we can query with SQL, such as Snowflake, BigQuery, Databricks, PostgreSQL, SQL Server or Oracle, including dbt projects. We only need read-only access.

What happens after the Assessment?

The roadmap is yours, written so your team can execute it without us. A month later we hold a 45-minute check-in, included in the price. No lock-in, no follow-up contract required.

What if we have several AI use cases?

We start with the one that matters most to your business. Each additional use case can be assessed separately.

How large can the use case be?

Typically up to 5 critical data sources: the tables, databases or pipelines your AI use case depends on. If yours needs more, we tell you before we quote.

What if our use case relies on documents (RAG)?

Today the Assessment covers structured data: warehouses, databases and the pipelines that feed them. If your use case relies on documents, tell us in the form and we’ll tell you honestly whether it fits.

What if the deliverables aren’t what we agreed?

Our scope guarantee covers it: if the deliverables don’t match the agreed scope, we keep working until they do, at no extra cost.

Which languages do you work in?

English, French and Spanish. The whole Assessment, report included, can be delivered in any of them.

Find out what’s holding your AI back.

Tell us about your AI use case. We’ll review it and reply within 2 business days, and tell you whether the Assessment is the right next step.

It takes 5 minutes and commits you to nothing. Every Assessment comes with our scope guarantee.

What should your AI use case do?
Where is it today?
When do you need an answer?
What’s getting in the way?
Tell us about your company
Where should we reply?
Press Enter to continue