Don't pay for AI strategy. Pay for AI implementation.
Most consultancies are still selling decks. The deck is the cheapest part. The build is what costs money. Don't buy them in the wrong order.
We see this every week. A founder pays $40k–$200k for an 'AI strategy' from a big-four or a tier-1 consultancy. They get a 90-page deck. The deck recommends 'transformation across 6 pillars'. Six months later, nothing's been built. The deck is in a Drive folder somewhere. The leak is still leaking.
Why this happens
Because strategy is high-margin, low-risk, and easy to scope. Implementation is the opposite. The big consultancies are good at decks. They're not built to ship working software inside your business.
Most of them then sub the implementation to a partner — usually offshore, usually under-scoped — and the build fails for boring reasons (no one understood your business, the integrations broke, no one's maintaining it).
What to do instead
Skip the deck. Pick the highest-ROI build (use the 30-minute exercise from another field note). Hire someone who builds for a living, not someone who recommends builds for a living. Pay for the build and the run. Skip the strategy phase entirely.
If your AI vendor's only deliverable is a deck, you've hired the wrong vendor.
When strategy is worth it
Genuinely: if you're a $50M+ business with multiple business units and you need a 3-year roadmap across them, hire a tier-1 strategy firm. They're good at that. Under $50M, just build.