When Your CEO Says "We Need an AI Strategy" — Your First Three Moves
Three concrete actions to take in week one so you read as the adult in the room, not the bottleneck, when the AI strategy mandate lands
You've heard a version of this by now. Leadership comes out of a board meeting, a conference, a competitor's press release — and suddenly it's: "We need an AI strategy."
Nobody looks at the engineers. They look at you.
Here's the thing: that moment is not a trap. It's a positioning window. The manager who walks out of week one having framed the conversation — instead of waiting for someone to hand them a framework — is the one who owns the workstream. The one who freezes, forwards a McKinsey PDF, and says "let me loop back" is the one who gets cc'd instead of consulted.
Three moves. Week one. Here's what they are.
Move 1: Name the problem before anyone names the solution
The first instinct — and the wrong one — is to start researching AI tools. Don't.
The first thing you do is write one sentence that says what business problem this strategy is supposed to solve. Not "leverage AI for efficiency." That's a budget line item dressed up as a strategy. Something like: "Reduce the time our ops team spends on manual reporting so they can focus on analysis."
That sentence does two things. It stops the conversation from becoming a vendor parade. And it signals — to everyone in the room — that you understand strategy is about the problem, not the technology.
Bring it to the first meeting. You don't need a deck. You need one sentence.
Move 2: Map who already has skin in the game
AI "strategy" meetings attract three types of people: the vendor-excited (will forward you product demos), the wait-and-see (will agree with whoever spoke last), and the quietly threatened (will raise concerns without naming them).
Your job in week one is to know who's in each camp before you're in the room together.
Spend 20 minutes having informal conversations — not about AI, about their work. "What's the thing your team spends the most time on right now?" is a better question than "what do you think about AI?" You'll learn more, and you'll be the person who listened before they had opinions.
This is political intelligence, not technical knowledge. It is wildly underused in these conversations.
Move 3: Set the scope ceiling in writing
Scope creep in AI initiatives is nearly instantaneous. What starts as "help our team work smarter" becomes "automate the supply chain" by week three if nobody puts a fence around it.
Before week one ends: send a summary email to whoever handed you this mandate. One paragraph. Confirm your understanding of the business problem, the team scope, and a realistic timeline for a first recommendation. Ask for a one-sentence confirmation that you have it right.
You now have three things: a paper trail, an agreed problem definition, and a sponsor who is on record. When the scope tries to expand — and it will — you have something to point to.
The signal here
None of this requires you to know anything about machine learning, large language models, or which vendor is best. What it requires is the thing you're already paid for: figuring out what the actual problem is, who the stakeholders are, and how to keep a workstream from sprawling.
The technical layer comes later. Week one is yours.
Forward this to the one person who's about to get handed this. They'll thank you when they come out of week one looking like the adult in the room.
P.S. — If you want the action layer beyond week one — what to say at the stakeholder meeting, how to run the vendor evaluation, what to put in the plan — that's what The Edge is for. Try it here.
This was a free one. There's a new one every week.
What's signal in the AI noise — and the move to make about it. No hype, no vendor pitch, no link dump.
Aiden Vector is an AI-assisted publication; this content is produced by AI under human editorial direction.