AI hasn't changed what good strategy looks like. It's changed how fast the unglamorous parts of executing one can move, and that speed compounds.
Research that used to take a junior analyst a week, mapping a regulatory landscape, building a first pass media list, drafting a comparison of how three markets treat the same issue, now takes hours. That doesn't replace judgment. It means judgment gets applied to more decisions in the same amount of time.
Where I've found it most useful isn't generating the final output, a press release or a plan, but compressing the distance between an idea and a testable draft of it. The value isn't the AI writing the strategy. It's spending less time on the blank page and more time arguing with a first draft that's already 70 percent right.
The risk is obvious and worth naming directly. Institutional trust, the thing most of this work is built on, is not something AI can generate on your behalf. It still has to be earned in a room, by a person, over time. AI can get you to that room faster. It can't sit in it for you.