AI strategy for leadership teams: pilots, build versus buy and moats
How to run an AI pilot that tells you the truth, when to build versus buy, where AI pays off first, and why the model you pick is rarely your advantage.
The questions that decide whether AI pays off are rarely about which model is best. They are about whether the pilot actually worked, whether to build or buy, where to start, and what will still be an advantage in two years when everyone has the same models.
My short answers: run pilots that measure the tool and not the expert using it, start where checking the output is cheap, buy before you build, and assume the model is a commodity. The posts below make each of those arguments properly, with the evidence behind them.
If you are preparing an AI roadmap or a board discussion, start with the first four.
The posts, in reading order
-
Your best people are the worst judges of your AI pilot
A convincing AI demo often measures the expert running it, not the tool. Design the pilot so the model's real contribution shows, or you'll buy a demo and deploy a letdown. -
Nobody has benchmarked AI on 90% of the work your company does
Public AI benchmarks cover the work that is cheap to grade, not the work that pays your bills. Why the unmeasured majority is where private evidence becomes an advantage. -
Your backlog is the only benchmark that matters
Public coding benchmarks have decoupled from real work. The teams getting value from AI build a small evaluation from their own merged pull requests instead. -
Point your AI where the checking is cheap
Claude found a novel cryptographic attack in about a week; human experts needed close to a month to trust it. That gap, not model quality, decides where AI pays off. -
Boring technology is an AI strategy
A decade-old rule for choosing dull, predictable tools turns out to be the sharpest discipline for deciding where AI belongs in your product, and where it does not. -
Cheap to fork, costly to keep: AI redraws build versus buy
AI collapses the cost of maintaining software you didn't write, so part of the buy column becomes buildable. The catch is the forks you now have to govern. -
Your model was never your moat
Open-weight models now match frontier quality at a fifth of the cost. For anyone building on AI, that shifts where durable advantage has to come from. -
Why the smartest model can make your product worse
Frontier models keep setting records, yet what decides whether a real-time AI product works is the second your user waits, not the model's score. -
450 million seats, 7% adoption: the real Copilot story
Microsoft folded its consumer AI companion into the enterprise product. The adoption number behind that decision says more than the U-turn itself. -
Your real AI advantage is how fast your company learns
Most teams point AI at producing output. The larger gain is reaching real competence in a new domain faster, without mistaking borrowed answers for understanding. -
When your AI vendor ships your product
OpenAI's move into legal shows what a platform really does when it enters your category: it commoditizes one layer of your stack and leaves you the accountable one.