Every founder I meet in 2026 has an AI strategy.
Most of them are wrong.
Not because they're stupid. They're not. Most are sharper than I am. They're wrong because they started with the technology and worked backwards to a problem, instead of starting with the problem and working backwards to the technology.
That's the difference between an AI strategy that prints money and one that burns it.
The 3-question diagnostic
Before you spend another dollar on AI tooling, AI agents, AI consultants, AI anything — answer these three questions. Out loud. To someone else. In one sentence each.
Question 1: What painful, repeated, expensive workflow are we trying to compress?
Not "we want to use AI." Not "AI for marketing." Not "AI to improve customer experience." Those are not workflows. Those are hopes.
A workflow is specific: "Our paralegals spend 11 hours a week answering case-status calls." "Our brokers miss 38% of inbound calls after 7 PM." "Our agents spend 45 minutes per listing writing descriptions."
If you can't name the workflow in one sentence with a number attached, you don't have a strategy. You have a vibe.
Question 2: What does success look like in 90 days, measured in dollars or hours?
"Save 8 hours per week per broker." "Recover 12% of after-hours leads that previously hit voicemail." "Cut document-review time on commercial leases by 60%."
If you can't put a number on the win in 90 days, you'll never know if it worked. You'll spend money. You'll have meetings. You'll write decks. Nothing will improve in any measurable way and in 6 months you'll quietly cancel the AI project.
Every successful AI deployment I've shipped started with a number. Every failed one started with "let's explore AI."
Question 3: Who in the organization will be measurably better off — and who might resist?
This is the question that kills most AI projects, and almost nobody asks it upfront.
If your "AI strategy" makes the people who use it daily feel threatened, replaced, or surveilled — they will quietly sabotage it. Not maliciously. Just by not feeding it data, not following the workflow, not reporting bugs.
The successful deployments I've seen all share one trait: the people doing the work today became MORE valuable, not less, after AI shipped. Their job got easier. They billed at a higher rate because they handled more. They went home at 5 instead of 7.
If you can't tell me who specifically wins from the AI deployment in language that PERSON would agree with, you have a politics problem, not a technology problem. Fix that first.
The pattern I see in failed AI strategies
Founder reads a McKinsey report about AI. Founder gets excited. Founder buys a $40K consulting engagement. Consulting firm produces a 60-page strategy document. Document recommends "AI Center of Excellence." Founder hires a Head of AI. Head of AI spends 6 months figuring out the org chart. Nothing ships.
Two years later: zero revenue impact, six-figure spend, no learnings.
Meanwhile: the founder's competitor read the same McKinsey report, ignored it, looked at their own business, identified that their inbound calls were leaking 38% of leads, deployed an AI voice agent in 2 weeks for $300/month, and is closing 22% more deals 90 days later.
What I tell every consulting client
Start small. Pick the smallest, most measurable workflow you can find. Ship the AI for it. Measure for 60 days. Decide whether to expand based on actual data — not based on what's exciting on LinkedIn.
If you build AI tooling for your business with this approach, you'll outperform 80% of the founders chasing AI hype because you'll actually ship things that work.
The entrepreneurial take
AI in 2026 is exactly like the internet in 1998. There's a stupid amount of hype. There's a lot of dumb money chasing dumb projects. There are massive winners being made quietly by founders who ignore the hype and just ship one boring useful thing at a time.
You can be the founder who chases the headline trend and burns $400K over two years. Or you can be the founder who asks the 3 questions, picks the smallest pain point, and compounds wins for the next 24 months.
One of those founders is much, much richer in 2028. Decide which one you want to be.