How to Run an AI & Data Readiness Assessment in 4 Weeks

Key Takeaways

• Four weeks is enough if the scope is fixed on day one. Two weeks to assess what you have, two weeks to decide what to build first. Assessments that drift past a quarter usually never had a scope.
• Do the prep before Week 1. An executive sponsor, a stakeholder list, and data access granted in advance decide whether Week 1 is discovery or waiting.
• Score readiness on evidence. Profile the actual data sources and review the actual integrations. Interviews tell you where to look. They do not tell you what the data looks like.
• Rank use cases on value, feasibility, and risk, with risk as a gate. A high-value idea on data that is not ready goes on the roadmap behind its fix, never in front of it.
• End with a roadmap someone can start next week. Owners, sequence, investment, and one pilot-ready first initiative. A report without those is shelfware.

Many AI readiness assessments fail on the calendar before they fail on the findings. A team kicks off "discovery," schedules interviews around everyone's travel, waits three weeks for read access to the ERP, and delivers a findings deck at the end of the quarter. By then the use cases have changed, the sponsor has moved on, and the first build is another quarter away.

The cost of skipping the assessment is worse. Gartner predicts that through 2026, organizations will abandon 60% of AI projects that are not supported by AI-ready data. In the same survey of 1,203 data management leaders, 63% of organizations said they either lack the right data management practices for AI or are unsure whether they have them. Most companies need the assessment. Few can afford to spend a quarter on it.

A fixed four-week assessment fixes both problems. It is long enough to profile real data and short enough that the findings are still current when the roadmap lands.

This guide walks through how to run an AI and data readiness assessment in 4 weeks: what to set up before Week 1, what to do and produce in each week, how to score readiness and use cases, and the pitfalls that push an assessment past its deadline.

What Is an AI & Data Readiness Assessment?

An AI and data readiness assessment is a time-boxed audit of the data, systems, workflows, people, and controls an AI program depends on. It produces a scored readiness baseline, a ranked list of AI use cases, and a roadmap that says what to build first and what has to be fixed before the rest can follow.

It answers two questions in order. First: what is true about our data and systems today? Second: given that, which AI use cases are worth building, and in what sequence?

An assessment is narrower than a full AI strategy engagement and later than technical due diligence on a deal. Strategy sets direction across the business. Diligence tests a target's claims before close. The readiness assessment sits between them: it tells an operating company what it can build next, with the data it already has.

What does a 4-week AI readiness assessment produce?

Each deliverable answers one question for one audience. If a deliverable has no reader, cut it.

Tip: See what a finished assessment covers, from the data readiness audit to the prioritized roadmap, on our AI & Data Readiness Assessment capability page.

Why Should an AI Readiness Assessment Take 4 Weeks?

Four weeks forces decisions. Each week has one job and one output, and the next week cannot start without it. The time box also limits the damage of a bad answer: if the data is worse than expected, you learn it in Week 2, not Month 3.

The table compares the three common lengths.

The four weeks split into two halves. Weeks 1 and 2 assess what you have. Weeks 3 and 4 decide what to build first.

What Do You Need in Place Before Week 1 of an AI Readiness Assessment?

Most four-week assessments that run long lose the time before they start. Settle these five things in the week before kickoff.

• An executive sponsor with authority to decide. Someone who can approve access requests in a day and who will own the roadmap at the end.
• A fixed scope. One business unit, one function, or one portfolio company. Write down what is out of scope, too.
• A stakeholder list with booked time. Process owners, data and IT leads, security or compliance, and two or three people who do the work every day. Book their interviews before Week 1 begins.
• Read access to the systems in scope. Request credentials, exports, or sandbox copies now. Access requests are the most common reason Week 2 slips.
• A definition of success. Two or three business outcomes the sponsor cares about, such as hours returned to a team, cycle time on a process, or error rates. Week 3 scores every use case against them.

Week 1: How Do You Run Discovery in an AI Readiness Assessment?

Week 1 is about how the work really runs. The output is a workflow map and an inventory of every data source and system in scope.

Interview stakeholders with a fixed question set

Use the same questions for every interview so the answers compare. Ask what the person does each week, which steps take the longest, which decisions they make by judgment, where the data comes from, and what breaks. Keep each interview to about an hour.

Watch the workflow, not just the description

Have the people who do the work walk through it on screen. You will see the spreadsheet that bridges two systems, the copy-and-paste step nobody mentioned, and the manual check that exists because a report was wrong once. Those details decide which use cases are real.

Build the data and systems inventory

For each source, record what it holds, where it lives, who owns it, how often it updates, and how you get access. Include the unofficial sources: shared drives, inboxes, and spreadsheets often hold the data an AI use case needs.

Week 2: How Do You Score Data, Systems, and Governance Readiness?

Week 2 turns Week 1's notes into a scored baseline. The rule is evidence over opinion: profile the data itself, test the integrations, and read the policies.

Score each dimension on the same scale, such as Ready, Developing, or Gaps found, and attach the evidence and a recommended fix to every score below Ready.

Test data against the use case, not in general

Data is never ready in the abstract. Gartner's guidance in the same release is that AI-ready data must be representative of the use case, including its patterns, errors, and outliers. Profile the sources that the candidate use cases need, and judge them against those use cases.

Use an existing framework for governance

You do not need to invent a governance standard. The NIST AI Risk Management Framework is intended for voluntary use and gives you a common vocabulary for AI risk that security, compliance, and business leaders can all work from. Score your current controls against it and list the gaps.

Tip: If Week 2 surfaces gaps in access controls or AI policy, our AI Security, Compliance & Governance capability covers how to close them before the first build.

Week 3: How Do You Score and Prioritize AI Use Cases?

Week 3 converts what you learned into choices. Start with a long list of use cases from the interviews, then score each one on the same three criteria.

Plot value against feasibility. The use cases in the high-value, high-feasibility box are your pilot candidates. High-value ideas with low feasibility go on the roadmap behind the fix that would make them feasible.

Treat risk as a gate

Risk works best as a gate on the ranking. A high-risk use case can still be worth doing, but it enters the roadmap only with a named control attached, such as human approval on every output or restricted data access.

Model ROI as ranges

For each top candidate, estimate the time or cost it returns as a low and high range, based on the baselines from Week 1. Ranges keep the conversation honest and show the sponsor where an assumption drives the answer.

Tip: If your use case list leans on how well people already use AI, our post on AI proficiency levels in private equity operations gives you a four-level benchmark for where teams stand today.

Week 4: How Do You Turn Readiness Findings Into an AI Roadmap?

Week 4 produces the roadmap and the scope for the first pilot. The test for every item: could a team start on it next week?

Sequence by dependency

Order initiatives by what each one needs. A data fix that unblocks three use cases goes first. A pilot that depends on an integration waits for it.

Name an owner and an investment for every item

Each roadmap item needs an owner, an architecture sketch, and an investment range. Without them, the roadmap is a list of wishes.

Scope the first pilot

Write the first pilot's scope as if you were going to start it on Monday: the workflow, the data sources, the people who will use it, the success metric and its baseline, and the review step for outputs. Present it with the roadmap so the sponsor approves both together.

Tip: In PressW's AI Transformation engagements, the four weeks of assessment and strategy lead straight into an Activate phase in weeks five through eight, where the first pilot is deployed.

How Do Strong Teams Keep an AI Readiness Assessment on Schedule?

Teams that finish in four weeks fix the scope early, chase access before kickoff, and judge every finding by whether it changes the roadmap. Teams that run long hit the same few problems.

Common pitfalls

• Starting without access. If credentials arrive in Week 3, Week 2 never happened. Request access in the prep week.
• Interview-only scoring. Readiness scored from what people say about their data is a guess. Profile the sources.
• Scope that grows each week. Every new business unit added mid-assessment costs a week. Log new ideas for the next round.
• Ranking on value alone. The most exciting use case is often the least feasible. Score feasibility and risk with the same care.
• A roadmap with no owners. Items without a named owner do not start.
• No first pilot. An assessment that ends without a scoped pilot leaves the team to restart planning, and momentum is gone.

Checklist for Evaluating Your AI Readiness Assessment Plan

Use this before kickoff, whether your team runs the assessment or a partner does.

• An executive sponsor is named and has approved the scope
• The scope is written down, including what is out of scope
• Stakeholder interviews are booked for Week 1
• Read access to every system in scope is requested or granted
• Two or three success outcomes are defined with a current baseline
• The readiness dimensions and scoring scale are agreed in advance
• Use cases will be scored on value, feasibility, and risk
• Each roadmap item will have an owner, sequence, and investment range
• The final readout is booked for the end of Week 4
• The team that will build the first pilot is known

Run Your AI & Data Readiness Assessment with PressW

A four-week AI and data readiness assessment works when the prep is done before Week 1, readiness is scored on real evidence, use cases are ranked with risk as a gate, and the roadmap ends with a pilot someone can start.

PressW runs AI & Data Readiness Assessments on the same shape as this guide. In our AI Transformation engagements, weeks one and two assess data, workflows, infrastructure, and team capabilities, and weeks three and four deliver a prioritized roadmap with owners, architecture, an ROI model, and investment requirements. The engineers who score your readiness are the ones who build what comes next: PressW has shipped 90+ production AI systems across 7 industries. For investment firms, we applied our 5 Pillars of AI Competency framework for Ecliptic Capital and delivered a technical report that informed their final investment decision.

If you want to know what four weeks would look like for your business or portfolio company, we will walk through the scope with you. Book a Discovery Call.

FAQ