A strategy is a pattern of choices

Strategy is visible in what you fund, what you refuse, what you measure, and what you stop. A useful AI strategy therefore cannot be a statement that AI is important. It must describe a pattern of choices:

• We begin with measurable business friction.
• We work one opportunity at a time until our decision system is reliable.
• We require an owner and users before we require a vendor.
• We run small experiments before making large commitments.
• We separate the decision to act from the decision to buy, assemble, or build.
• We treat adoption as part of the design.
• We stop work when the evidence does not earn further investment.
• We govern the portfolio in business terms.

That is enough strategy to begin. It tells the company what earns attention before any tool earns a budget.