You’ve probably been told you need to “do AI.” You’ve probably also watched a peer waste $200K on a pilot that died on a shelf. The numbers aren’t with you: more than 80% of AI projects fail, according to RAND — twice the failure rate of IT projects that don’t involve AI. And it’s getting worse, not better: the share of companies abandoning most of their AI initiatives jumped to 42% in 2025, up from 17% the year before, per S&P Global Market Intelligence.
But the failures aren’t because AI doesn’t work. AI works. Claude works. ChatGPT works. The failures happen because the business wasn’t ready — the data was a mess, the team wasn’t aligned, no one owned the rollout, the right problem wasn’t picked, or leadership didn’t commit past month three.
The good news: readiness is diagnosable. There are specific questions that, if you can answer “yes” or with a clear plan, predict success. If you can’t, they predict failure.
These are the 15 we use with prospects on a first call. Run them on your own business. If you score well, you’re ready to invest. If you don’t, fix the foundation before you spend.
Five dimensions, three questions each
Dimension 1: Strategic clarity
1. Can you name, in one sentence, a specific business problem that AI is going to solve?
Not “we should explore AI.” Not “we need to be more data-driven.” A specific, painful, measurable problem. Example: “Our quoting takes 5 days; we’re losing 15% of bids to faster competitors.”
If you can’t, you’re shopping for technology in search of a problem. Don’t buy yet.
2. Can you name the top 3 places AI could most impact your business in the next 12 months?
If you can name 3 and they’re prioritized, you’re ready. If you can name 3 but they’re scattered (“legal, ops, marketing — all of them”), you have an inventory but no plan. If you can’t name any, you have a wish, not a strategy.
3. Who in your organization owns AI strategy?
Acceptable answers: the CEO with a dedicated owner reporting in; a specific operations or business leader; a real (not theoretical) committee.
Unacceptable answers: nobody, IT, “we’ll figure that out as we go.” If you can’t name the owner, the AI project will stall in month four when someone has to make a hard call.
Dimension 2: Operational readiness
4. How would you describe the documentation of your core operational workflows?
If your workflows are mostly tribal knowledge in people’s heads, AI can’t automate them. You need workflow documentation before AI can sit on top.
This isn’t a deal-breaker — sometimes the first AI project IS documenting the workflows. But know that’s the project.
5. How accessible and structured is your company’s data?
Same point. If your data is in 6 systems that don’t talk, AI surfaces every inconsistency. Some companies need a data-readiness sprint before any AI work. That’s fine. It’s better than building AI on a foundation that won’t hold.
6. What’s your team’s relationship with new technology?
Resistant? Mixed? Open but cautious? Genuinely curious?
This determines how much change management your AI rollout needs. A team that adopts new tools willingly will get to value 6 months faster than a team that resists. Both can succeed — but they need different rollout plans.
Dimension 3: Use case maturity
7. How many AI use cases are currently in production at your company?
In production = actively used + producing measurable value. Not “we have ChatGPT licenses.” Not “our marketing team uses it sometimes.”
- If you have zero: you’re early. That’s fine — most companies are.
- If you have 1–2: you’ve started learning. Now you can scale.
- If you have 3+: you’re past curious. The next question is whether you have a structured roadmap or random experimentation.
8. What’s your largest AI win to date?
The honest answer matters. “Someone using ChatGPT for emails” is a different starting point from “a documented cross-functional process improvement with measurable ROI.”
If your largest win is individual productivity, that’s normal. Don’t pretend it’s bigger than it is.
9. Describe one workflow at your company that AI should be able to improve but currently isn’t.
If you can name one specifically, you have a real opportunity to pursue. If you can’t, you don’t have enough operational understanding to invest yet — you’re being sold to, not buying.
Dimension 4: Investment & capacity
10. What’s your company’s annual revenue and EBITDA range?
This shapes what’s realistic. AI Office partnerships at $2,500–$10,000/month fit lower mid-market service businesses ($5M–$100M revenue, $1M–$10M EBITDA). Smaller, you can’t absorb the change. Larger, you usually have internal capability already.
11. How would you characterize your appetite to invest in AI in the next 12 months?
Minimal, exploratory, active (budget allocated), aggressive (priority), or board-level strategic?
The right answer for most mid-market operators is active — budget allocated, looking for the right partner. Aggressive without clarity wastes money; minimal misses the window.
12. Who would lead an AI initiative if you launched one today?
The honest answer is usually “nobody right now” or “an existing leader as a side responsibility.” That’s okay — but it means you need a partner who’ll fill the seat with you, not a vendor expecting you to drive.
If you say “someone would have to be hired,” you’re 6 months away from being able to start internally. A partner can bridge that gap.
Dimension 5: Risk & governance
13. How would you describe your AI governance?
None, informal verbal guidance, basic policy that isn’t enforced, documented policies with training, or mature governance with audit and monitoring?
Most lower mid-market businesses are at “informal” or “basic policy” — that’s fine for the first wave of AI work. You don’t need enterprise-grade governance to deploy a quoting assistant. You do need it before you give an agent authority to send money.
14. What’s your biggest concern about deploying AI more aggressively?
Data privacy/security? Accuracy/trust? Employee resistance? Cost vs. value? Regulatory exposure?
There’s no wrong answer. The wrong move is not knowing what your biggest concern is — that means you haven’t thought hard enough about it yet.
15. If your AI deployment got something visibly wrong in front of a customer, what’s your plan?
This is the question most companies don’t answer until it happens. The right answer is a documented escalation path with human-in-the-loop checkpoints where the consequences matter. The wrong answer is “we’ll deal with it when it comes up.”
What to do with your score
If you scored well on Dimensions 1, 4, and 5, and have at least a plan for 2 and 3 — start. An AI Office partnership starting at $2,500/month is designed for exactly your situation.
If you scored poorly on Dimension 2 (operational readiness), the first AI project should be a data-and-workflow readiness sprint, not an AI build. That’s a $25K sprint, not a $250K program, and it pays for itself.
If you scored poorly on Dimension 3 (use case maturity), you need an opportunity discovery process before any build. That’s literally what an AI Office strategist does in months 1–2 of an engagement.
If you scored poorly on Dimension 1 (strategic clarity), don’t spend money on AI yet. Spend an afternoon with your leadership team getting clear on the problem first.
Want a personalized read?
We built a 10-minute AI Readiness Assessment that produces a custom 2–3 page diagnostic: your score across all five dimensions, your top 3 priorities for the next 90 days, and a recommended path forward. It’s free. Not a sales tool — a real diagnostic.
If your score lands in the “ready” range and you want to talk about how an AI partnership would look in practice, we can do that in 30 minutes.