An AI assessment is a short, honest look at where AI would save you real time and money, and where it wouldn’t. Not a maturity score. Not a tech checklist. Just a map of your work with a simple question next to each part: should AI touch this, or not?
The outcome should be three clear priorities you can act on this month. Not a 47-page report. Not a level on a ladder. Three things to try first, ranked by how much time they’d buy you back.
That sounds simple, and it is. But most businesses skip this step entirely. (If you already know AI fits and want to score how ready your team is, the AI readiness assessment goes deeper. This post is the step before that.) Cisco surveyed 7,985 leaders across 30 countries and found 98% feel urgency to use AI. Only 13% are actually ready. That gap exists because people jump to tools before figuring out where those tools belong.
Small businesses feel it even more. The NFIB found that only 24% of small employers currently use AI, but 63% think it’ll be important within five years. The SBA tracked a 40% jump in small business AI adoption in just six months. The window to figure out where AI fits your work is right now.
Why most AI assessments miss the point
Most AI assessments you’ll find online use something called a maturity model. That’s a scorecard that rates your company on a scale of 1 to 5 across categories like “data readiness” and “governance.” You answer some questions, you get a number, and then you’re told to climb to the next level.
The problem: a score tells you where you stand on an abstract ladder. It doesn’t tell you what to do Monday morning.
Barry O’Reilly, author of Unlearn and Lean Enterprise, puts it plainly: maturity models “create the appearance of progress without guaranteeing impact.” You can score well on governance and data strategy while your team still pastes text into ChatGPT one message at a time.
These models were built for big companies with dedicated AI teams. A peer-reviewed study of AI maturity frameworks found they have “limited empirical grounding” in small-business settings.
Most assume you have departments, data infrastructure, and a formal governance program. If you’re a 5-person team, the whole exercise is a category error.
My take: I’ve seen founders spend a week on a maturity quiz and come away knowing they’re “Stage 2.” Great. What do they do with that? Nothing, because the score doesn’t connect to their actual work. The useful question is never “what stage are we?” It’s “which three tasks should AI handle first?”
The data backs this up. RAND found that about 80% of AI projects fail. Not because of the technology. Because of scattered data, unclear goals, and missing skills. A maturity quiz wouldn’t have caught any of that. An honest look at how the work actually gets done would have.
And the failures keep coming. S&P Global reported that 42% of companies scrapped at least one AI initiative in 2025, up from 17% the year before. The ones that survived started smaller and more honestly.
Ethan Mollick, a Wharton professor who led the Harvard/BCG study of 758 knowledge workers, has a name for this problem. He calls AI’s capability map “jagged.” AI is superhuman at some tasks and terrible at others that look equally hard from the outside.
You can’t predict which side your specific workflows fall on by filling out a quiz. You have to try.
Klarna learned this the expensive way. They replaced 700 customer service agents with AI in 2023, called it a win, then reversed course in 2025 and started rehiring. The CEO admitted: “We focused too much on efficiency and cost.”
AI handled simple queries fine. It fell apart on anything complex, emotional, or judgment-dependent. The right question was never “can AI do this job?” It was “which parts?”
The three things a useful AI assessment checks
Forget the 24-question quiz. A useful AI assessment for a small team checks three things.
1. Where your time actually goes.
Not your ideal process. Your real one. Track your work for a week. Where do you spend the most hours? Where is the work repetitive, predictable, or copy-paste? Be honest. Most of us spend more time on admin than we’d like to admit.
2. Where AI fits (and where it doesn’t).
Score each workflow on four things:
- How often you do it (volume)
- Whether the steps are the same every time (consistency)
- Whether clean inputs exist (data quality)
- What happens if AI gets it wrong (stakes)
High volume + high consistency + low stakes = your first AI candidate.
Low volume + high stakes + lots of judgment = leave it alone.
My take: The biggest trap I see is people trying AI on the thing they want help with most, instead of the thing AI is actually good at. Your quarterly strategy deck needs judgment and context. Your daily social media captions need a format, a tone, and a deadline. Guess which one AI handles better.
3. What’s actually blocking you.
Usually not money. Usually not tools. Usually one of three things:
- Skills. Your team doesn’t know how to use the tools well. EY surveyed 15,000 employees and found 88% use AI at work, but only 5% use it in ways that actually change how they work. Only 12% get enough training. Skills are the quiet blocker.
- Data. Your information is scattered across systems. Gartner predicts 60% of AI projects without clean, ready data will be abandoned by 2026. If your customer data lives in three spreadsheets and a CRM that nobody updates, AI can’t do much with it. (If that sounds familiar, AI data solutions is a good place to start.)
- Trust. Your team doesn’t trust AI output, so they redo everything by hand. That’s not irrational. It just means you need to start where the stakes are low so trust can build naturally.
Understanding your real barriers to AI adoption is half the assessment.
McKinsey’s State of AI report found that 78% of companies now use AI somewhere. But only about 6% get real financial return from it. The gap isn’t tools. It’s execution.
BCG found that 60% of companies don’t even track whether their AI investments are paying off. They literally don’t measure it.
An assessment that doesn’t ask “how will we know this worked?” is incomplete.
How to run your own AI assessment in a week
You don’t need a consultant or a tool for this. The workflow audit method is simple: inventory, track, score, pick. You can do it in one focused week.
Day 1-2: List every repeatable task.
Open a spreadsheet. Write down every task you or your team does regularly. Who does it, how often, with what tool. Include the boring stuff. Especially the boring stuff.
If you want a head start, grab an AI checklist to make sure you’re not missing any categories.
Day 3-4: Track real time.
Next to each task, write down how long it actually takes. Not how long it should take. How long it takes. Note which tasks follow the same steps every time and which ones need judgment calls.
Day 5: Score each task.
Multiply: frequency x time x consistency. Divide by stakes. Sort highest to lowest. The AI for small business marketing workflows usually cluster near the top: social posts, email drafts, reporting, customer replies.
A real example: writing social media captions. Daily, 30 minutes, same format every time, low stakes if one caption is a bit off. That scores high. Responding to an upset customer about a billing error? Low frequency, high stakes, needs judgment. That scores low.
Weekend: Pick the top 3.
These are your AI priorities. Not 15 use cases. Not a roadmap. Not a vision deck. Three things to try first, starting next week.
Practical tip: if a task takes less than 15 minutes and happens less than once a week, skip it. The juice isn’t worth the squeeze.
What to do with your three priorities
For each priority, find the simplest tool that fits. Check the best AI tools for business if you’re not sure where to start. Or look at the best AI tools for marketers if your priorities are marketing-heavy.
Then start with one. Not all three. Run it for two weeks. Measure one thing: hours bought back per week. Not “AI maturity.” Not “adoption percentage.” Hours.
Barry O’Reilly suggests tracking two numbers: “minutes saved per week” and “decisions improved with AI assistance.” That’s it. Mastery through practice, not stage-climbing.
After two weeks, if it’s working, add the second priority. If it’s not, figure out why. Usually it’s one of those three blockers from earlier: skills, data, or trust. Fix the blocker, then try again.
When to bring in help: if your priorities involve scattered data, cross-team workflows, or areas where mistakes cost real money. That’s when a guided assessment pays for itself. An AI strategist can spot things you’d miss on your own. And an AI adoption framework gives you the structure to roll them out without chaos.
If you want the full step-by-step, there’s a good walkthrough on implementing AI. And if you’re weighing outside help, here’s what AI consulting for small businesses actually looks like.
How I can help
If you’ve read this far and thought “I could do this myself,” you probably can. That’s the point. A spreadsheet and a week of honest tracking gets most small teams 80% of the way there.
But if your situation is messier? Data in five systems, a team that’s skeptical, a board asking for an AI plan by next quarter. Sometimes it helps to have someone who’s done this before sit with you and work through it.
That’s what I do. We map your workflows, find the real priorities, and build a plan you can start on. No maturity scores. No 50-page decks. If that sounds useful, let’s talk.
FAQ
What is an AI assessment?
A business AI assessment maps where AI would save you real time and where it wouldn’t. It looks at your actual workflows, scores them by volume, consistency, and stakes, and gives you a short list of priorities. The goal is action, not a grade. For a deeper readiness check with scoring, see the AI readiness assessment.
How do you assess AI readiness?
Track your workflows for a week. Score them on how often they happen, how repetitive they are, and what happens if AI makes a mistake. The highest-scoring tasks are your starting point. For a structured scoring system, the AI readiness assessment walks through five dimensions: workflows, skills, data, tools, and culture.
How much does an AI assessment cost?
You can run a basic one yourself in a week with a spreadsheet. It costs nothing but time. Professional assessments from consultants typically range from a few thousand to tens of thousands of dollars, depending on how many teams and systems are involved. For most small teams, the DIY version is enough to get started.
What’s the difference between an AI assessment and an AI audit?
An assessment is forward-looking: where should we use AI? An audit is backward-looking: is our current AI use working? You’d typically assess first, adopt, then hire an AI auditor later to check whether your AI tools are delivering and being used responsibly.
Do I need an AI assessment tool?
Not to start. A spreadsheet and honest time-tracking cover most small teams. Tools help when you’re running the assessment across multiple departments or want to benchmark against industry data. But don’t let “finding the right tool” become the reason you never start. The assessment itself is the value, not the software.