What to Look for in an AI Readiness Assessment

What a readiness assessment should answer

At its core, a readiness assessment is trying to answer a straightforward question: where is your organization today, and what would need to be true for AI to genuinely help your work?

A good assessment should leave you with clear answers to five things:

1. Where AI could help — specific workflows and pain points where the technology fits, not a generic list of possibilities
2. Where your data stands — whether you have the information infrastructure to support the tools you're considering
3. How your team feels about it — where there's enthusiasm, where there's anxiety, and what skills would need to develop
4. What risks you need to manage — privacy concerns, ethical considerations, and compliance obligations specific to your work
5. What to do first — a realistic sequence of steps sized to your budget and capacity

If the assessment you're considering doesn't touch all five of these areas, it's probably too narrow. And if it touches all five but produces a 60-page report, it's probably too broad for where most organizations are right now.

What the process typically looks like

The assessments I run tend to follow a similar arc. They start with conversations, move into observation, and end with recommendations. The whole process usually takes two to three weeks.

Staff interviews are where most of the value comes from. I talk with people across different roles and departments to understand their actual daily work, not just what's in their job description. The person processing intake forms has a very different perspective on where time gets wasted than the person writing grant reports, and both perspectives matter.

These conversations also surface how people feel about AI in general. Some staff are already experimenting with tools on their own, some are curious but cautious, and some are worried about their roles changing. Understanding that landscape is essential because the best AI implementation plan in the world fails if the people who need to use the tools don't trust the process.

Workflow review means looking at how information moves through your organization. Where do people re-enter data? Where are they copying and pasting between systems? Where does a task require four steps that could require one? These are the places where AI tends to create the most immediate value, and they're often invisible to leadership because the people doing the work have long since adapted to the friction.

A half-day workshop brings the team together to look at what the interviews and workflow review surfaced. This is partly educational and partly collaborative. The staff who will use these tools should have a say in which problems get prioritized, and a workshop creates the space for that conversation in a structured way.

What you should get at the end

The deliverable from a readiness assessment should be something you can act on, share with your board, and reference when writing grant applications. At a minimum, you should expect:

• A prioritized list of opportunities, ranked by impact and feasibility for your specific organization
• An honest evaluation of your data infrastructure and what gaps you'd need to address
• A summary of staff readiness, including training needs and concerns that should be addressed
• Specific tool recommendations with cost estimates, not just categories of technology
• A phased implementation roadmap that respects your budget and your team's capacity to absorb change

The document should be written in language your board can understand, because many organizations use the assessment to make the case for investment in AI capacity. If your assessment report reads like it was written for a CTO, it's not serving its audience well.

Red flags to watch for

Tool-first recommendations. Be cautious of any assessment that arrives with a specific product already in mind. A good assessment starts with your problems and works toward solutions, and the solutions might include tools you've never heard of or approaches that don't involve AI at all. If the assessor has a partnership or reseller relationship with a particular vendor, that's worth knowing upfront.

No staff contact. An assessment that only talks to leadership misses the most important information. The people doing the work know where the real bottlenecks are, and they're the ones who will determine whether any new tools get adopted. If the proposal doesn't include interviews with frontline staff, ask why.

Vague timelines and deliverables. Before signing anything, you should know exactly what you'll receive, when you'll receive it, and in what format. "A comprehensive report" could mean anything from five focused pages to a binder nobody reads. Ask for a sample deliverable or a detailed outline of what the final document will include.

No mention of risk or policy. Any assessment worth paying for should address the governance side of AI adoption. If the proposal focuses entirely on opportunity and efficiency gains without discussing data privacy, ethical guardrails, and disclosure requirements, it's incomplete.

Do you need one?

An honest answer: not every organization does, at least not yet. If your organization has fewer than ten staff and you're mostly wondering whether ChatGPT could help with your newsletter, a full assessment is probably more than you need. A short consulting conversation and an AI use policy might be a better starting point.

An assessment makes the most sense when you have multiple departments, enough staff that adoption requires coordination, and either a board or a funder asking you to be deliberate about how you approach AI. It's also valuable when you're considering a meaningful investment in new tools and want to make sure you're solving the right problems before committing budget.

The goal is always the same: to close the gap between what your team is doing today and what would help them do their work better, with a plan that feels achievable rather than overwhelming.