What an AI readiness assessment should actually cover

Most AI readiness assessments produce a maturity score and a slide deck. That is rarely what a business needs. A useful assessment answers a narrower question: which one or two workflows are ready for AI now, what has to be true for them to work, and what should wait.

Readiness is mostly a data and workflow question

In February 2025 Gartner predicted that through 2026 organizations will abandon 60% of AI projects that are not supported by AI-ready data. In the same release, Gartner reported that in a survey of 1,203 data management leaders, 63% said their organizations either do not have, or are unsure they have, the right data management practices for AI. McKinsey's 2025 global survey found that 88% of respondents' organizations use AI in at least one business function, but only 7% say AI is fully scaled, and the survey points to redesigning workflows as a key factor among the companies getting the most value. In other words, using AI is common; making it work inside real operations is the hard part, and it depends on data and workflow more than on the model.

The five areas a real assessment covers

• Workflows: a list of the 10 to 20 recurring workflows, who performs them, how often, how long each item takes and what an error costs. Candidates are ranked by volume and pain, not by how interesting the technology is.
• Data: where the inputs actually live (inboxes, PDFs, spreadsheets, the ERP or CRM), whether they are consistent, who owns them and who is allowed to touch them.
• Systems and integration: which systems can be read through an API, a database connection or a scheduled export, which can be written to, and which steps must stay manual.
• Risk and governance: who approves AI output before it reaches a customer, what customer or employee data is involved, and how mistakes are caught. The NIST AI Risk Management Framework is a useful shared vocabulary here; it organizes the work into four functions called Govern, Map, Measure and Manage.
• People and change capacity: who owns the result, who will use it every day, and how much time they have to review and correct it in the first month.

What you should receive at the end

A good assessment ends with a ranked shortlist, not a score. For each candidate workflow it should state the baseline measure (for example, hours per week or days to respond), the named owner, the systems it touches, the main risk and the safeguard, and a clear go, wait or no-go. It should also include a short list of things not to do yet, with reasons.

Red flags

• It never names a specific workflow in your business.
• It gives a score but no baseline you could measure against later.
• It recommends a product before anyone has looked at your data.
• It has no "do not do this" list.

If you want a second opinion on an assessment you already have, or a scoped one for your operation, you can reach us through the contact page.

Sources

• Gartner, "Lack of AI-Ready Data Puts AI Projects at Risk" (Feb 26, 2025)
• McKinsey, "The State of AI: Global Survey 2025"
• NIST, Artificial Intelligence Risk Management Framework (AI RMF 1.0)