Conversation 3: On AI Strategy

Leading Through Hard Conversations

Conversation 3: On AI Strategy

The question you will hear: "What's our AI strategy?"

Your stance: AI strategy and engineering strategy are the same document. Same conversation. Same investment thesis.

The wrong answer is a deck full of tools being evaluated. Every technology organization is evaluating the same tools. That is shopping, not strategy, and boards recognize it immediately.

How to lead in that room: Bring three things.

  1. Evidence of leverage today — the organization's own numbers, not vendor claims. Cycle time, review throughput, incident resolution, experiment velocity.
  2. Where investment is directed next — not promises of outcomes, but clear learning goals for the next two quarters with signals that will confirm or redirect the bet.
  3. A "not doing" list — the boundary between strategy and wish list is the willingness to say no. If you cannot name what you are explicitly not doing, you do not have a strategy. You have a backlog.
Before After
AI strategy in a separate document AI strategy embedded in every engineering and product decision
Boards receive vendor evaluation decks Boards receive internal metrics, investment theses, and "not doing" lists
AI treated as a procurement decision AI treated as operational transformation requiring hiring, training, process change
Strategy defined by tools to adopt Strategy defined by capabilities being built and problems chosen not to solve
Success measured by pilot launches Success measured by operational integration and measurable productivity shifts

What to avoid: A deck full of tools being evaluated. Promises of outcomes without defined learning signals. Letting the conversation become about AI as a separate initiative from engineering.