#005 - Strategy is for amateurs.
Strategy is for amateurs. Execution is for professionals.
Not quite.
You have the credentials. Great university. Real experience. You enjoy strategy and you’re good at it. Yet your plans stall. The deck is sharp, the meetings are smooth, but the results… thin. Easy to blame company culture, headquarters, or “the market.” But is that the full story?
Be honest. You felt the program could fade. You saw eight priorities jammed into one quarter. Still, you went along. Because one priority doesn’t look ambitious. It looks… small.
And yet small is often the only path to big.
The AI strategy trap
This is where AI strategy gets dangerous. The hype pulls you up and away. Grand visions replace operational truth. So your strategy is like a promise, not a plan.
Let's consider a different approach: an operating system for AI strategy that protects imagination while forcing execution. Imagine, Design, Run. Simple words, strict behavior.
“Run” is not a synonym for “we’ll try.” Run means shipping production-grade capabilities that change a P&L. Real users. Real data. Real risk. Real wins. If it does not ship, it does not count.
Most teams miss here. They do AI maturity checklists that obsess over data and ignore people. They run tiny pilots until everyone forgets why they started. They declare "scale" when four teams have access to a notebook. That is theater. You need a spine.
Execution that actually ships
Here is what execution looks like when you mean it:
-
Stop bullsh*tting yourself with corporate language. Run an AI readiness check that measures more than data. Consider culture, decision-making, incentives, and skills. If product teams need two weeks and three tickets to get a dataset, you are not ready. If managers reward “new ideas” but not shipped outcomes, you are not prepared. Name the blockers in plain words. Then remove them.
-
Select pilots that move a business metric. Two or three. Not ten. Tie each to one or two KPIs you already track. Cost per claim. On-time delivery. Lead time. Define clear guardrails and kill rules. No vanity demos. There is one exception. If you need a low-risk "learn the tooling" pilot, do it. Call it training. Do not confuse it with a business bet.
-
Prepare to scale while you pilot. Build the minimum platform that lets you scale success. Data for the chosen domains. Basic governance. Security patterns. A shared prompt and component library. Reusable integration paths. Think “starter kit,” not “cathedral.” You are laying track just ahead of the train.
-
Scale one winner to the max. Pick your best pilot and treat it like a product. Harden it. Put it under real load. Roll it out team by team, region by region. Train people. Update runbooks. Wire telemetry to your finance metrics. Fund it like a line item, not a science project. Only then start the next one.
That is the list. One list. Everything else is posture.
What it looks like in real work
What does this feel like in practice? A claims AI agent that shortens handling time gets pushed to every high-volume corridor, not stuck in the innovation lab. A replenishment model that reduces stockouts gets embedded in the weekly store rhythm, not just showcased at the town hall. An AI code assistant that improves code quality becomes part of the SDLC, with standards, reviews, and measurable gains. No fireworks. Just a disciplined rollout.
You will face friction. Legal will worry. Security will add steps. Ops will ask for "one more tweak." Good. That is the work. Respect the concerns, but keep the line moving. Your job is to balance speed and safety without losing momentum.
Courage in the decision. Calm in the process. Clarity in the numbers.
How do you know you are winning? Leaders stop asking for "the AI update" and start asking for the business trend that AI now drives. Teams pull the new capability into their own rituals. Finance sees movement in cash, cost, or growth without translation. The story gets boring. That is a great sign.
Choose the hill. Take the hill.
A quick note on strategy itself. You still need it. Strategy chooses the hill. Execution takes the hill. Without a choice, you drift. Without the climb, you brag. The best leaders do both with force and focus. They pick fewer hills and put more weight behind them. They say no more than they say yes. They understand that one strong win pays for five experiments.
If you want a framework, keep it light. Imagine the possibilities with your best people, close to real work. Design a small number of bets with explicit trade-offs. Then Run with a cadence the business can feel. That's the operating system. Nothing fancy. Just a way to protect outcomes from opinions.
Ask yourself three questions before your next “AI strategy” meeting. What single metric am I trying to change this quarter? What one pilot has the best path to production? What must I remove this week to make that path shorter?
If you cannot answer, you do not have a strategy. You have a wish.
Your move.
Read more in my newest blog article:
👉 Building AI Strategy Part 3: Run to Win.
