What I Learned from 50 AI Readiness Assessments

After fifty of these, the patterns are hard to miss. Here's what most businesses are getting wrong — and one opportunity almost everyone is sitting on.

What the Patterns Actually Look Like

When we started doing AI Readiness Assessments, I thought the patterns would vary a lot by industry. A landscaping company would have different issues than a law office. And in the details, that's true. But at the structural level, the patterns converge more than I expected.

The Most Commonly Missed Opportunity

In roughly 70% of assessments, the biggest missed opportunity is administrative communication — specifically, the time spent writing emails, drafting proposals, creating reports, and composing updates that follow a predictable structure every single time.

These tasks are invisible because they don't feel like the "real" work. But when I actually track time with clients, it's not unusual to find that a two-person sales team is spending six to eight hours per week on email drafting that follows a clear enough pattern for AI to assist with significantly. At $45/hour fully loaded, that's a meaningful number per week in potential time savings from one workflow change. People are surprised by that. They shouldn't be — they just never tracked it.

Common Misconceptions I Run Into

• "We're not technical enough for AI." The current generation of AI tools requires approximately the same technical skill as writing an email. This objection usually means "we haven't been shown a concrete use case for our work" — which is a solvable problem.
• "We need to solve our data problems first." Some AI tools require clean, well-structured data. Most practical tools for small businesses don't. You don't need a data warehouse to use an AI writing assistant.
• "Our industry is too regulated for AI." Sometimes this is true. More often, it means "we need to be careful about which tasks we use AI for" — which is different from avoiding it entirely. Most regulated businesses have plenty of internal tasks where AI is perfectly appropriate.

A Quick Self-Assessment

Spend 20 minutes on this before talking to anyone:

1. List your five most time-consuming recurring tasks — not projects, recurring tasks you do every week.
2. For each one, ask: does this task involve taking roughly similar inputs and producing roughly similar outputs? If yes, it's a candidate for AI assistance.
3. Ask: what would go wrong if this output had an error? High-stakes outputs need more human review. Lower-stakes first drafts have more room to experiment.
4. Count how many "yes" answers you got to step 2. Three or more means real opportunity. One or two still gives you a starting point.

The full assessment I do is more thorough — it covers technology stack, team dynamics, and workflow documentation. But this exercise will tell you whether you're in the "definitely worth exploring" category or the "sort out some fundamentals first" category. Most businesses land in the first one.