Thursday. The strategy offsite.
Someone clicks to the slide titled "Our AI Strategy."
Eight bullets. A copilot in the product. A model fine-tuned on support tickets. An "AI-first roadmap" by Q3.
The room nodded. The CEO looked relieved. Someone wrote "ship it" in the margin.
Everyone walked out feeling like a decision had been made.
Nothing was decided.
A capability was listed. A direction was approved based on gut-feeling. The hard part got skipped — and the relief in the room is exactly why nobody noticed.
AI is not your strategy.
It's the second time this decade you've handed your judgment to something else and called it one.
The first time, the proxy was the customer.
"Customer-obsessed." "Listen to the user." "Fall in love with the problem."
It sounded like rigor. Often it was abdication.
Because "the customer asked for it" is not a decision. It's a way to avoid making one.
You could fund almost anything and point to a customer who wanted it. The roadmap filled up. The trade-off never got named. And the word everyone used for this was discipline.
Now the proxy has changed costume.
"The model can do it." "Our competitors are AI-first." "The board wants an AI story."
Same move. A fashionable input stands in for the call you're paid to make.
Not because product leaders are lazy. Because judgment exposes you and a proxy protects you.
If the customer asked, it isn't your fault when it fails. If everyone is adding AI, you're not the one who looks exposed for adding it too. The proxy is a place to hide the decision — and the bigger the room, the more tempting the hiding place.
If you've done this long enough, you've watched the costume change before.
"Mobile-first" was a proxy once. "Cloud" had its turn. "Platform" did too.
Each one arrived as the thing every serious company simply had to have.
And each one let a room full of capable people skip the only question that mattered — whether this business, with this cap table, should pay for it now, and at the expense of what.
The ones who got it right didn't avoid the trend. They priced it. They named the gap it closed and what it would cost to follow it.
AI is the loudest costume yet. It is not the first. And the discipline it demands is exactly the same.
Strip the costume off and the real face is simple.
A strategy is a choice about where capital goes — and what you refuse to fund as a result.
AI is a capability you allocate against that choice. It is never the choice itself.
Confuse the two and you stop doing the only work that was ever yours: deciding which gap is worth closing, and what you'll starve to close it.
Watch what the costume hides.
AI carries a COGS line most "AI strategies" never put on the slide. Inference isn't free. Every call has a marginal cost that compounds with usage — the more customers love the feature, the more it bleeds.
So "add AI everywhere" isn't a roadmap. It's a gross-margin decision nobody priced.
Include it in the core product, or monetize it separately? That's a capital and pricing call — not a feature call. Get it wrong and you erode margin quietly for a year before anyone connects the slide to the P&L.
And the right answer changes with who funds you.
A VC-backed company near its next raise can spend on an AI bet that buys a growth narrative — the story is the asset.
A PE-backed company eighteen months from exit cannot — the margin hit shows up directly in the multiple.
A public company will answer for that COGS line on the very next earnings call.
Same AI feature. Three different correct decisions. The capability told you none of them.
Only judgment does that. Only someone who knows the cap table, the exit window, and the gap can say which version of the bet is right for this company, this quarter.
I watched a Series B product leader learn this the expensive way.
His team shipped an AI assistant because the competitor announced one. Great demo. The board loved it. Adoption climbed for two quarters.
Then the cloud bill arrived.
The feature he'd added to look innovative was now the single largest line item dragging gross margin. He hadn't decided to spend that money. He'd matched a competitor's press release and let the COGS find him later.
He didn't have a product problem. He had a decision he never actually made.
That's the real cost of the proxy.
You ship AI because it was on the slide. Usage climbs. Margin slips. The feature that was supposed to signal innovation starts dragging the business.
And the board stops asking about your AI story.
They start asking why gross margin fell four points — and you don't have an answer, because you never made a call. You ratified a trend.
The companies that win this cycle won't be the ones who added AI fastest.
They'll be the ones who knew which gap they were closing, what it cost to close it, and what they chose not to fund to pay for it.
So the question on the slide is the wrong one.
From: "What's our AI strategy?"
To: "Which business gap are we closing, what will the capability cost to run, and what are we not funding to pay for it?"
The first question has a thousand confident answers.
The second has one honest one.
Here's the test to run before the next "AI-first" slide goes up:
Would this bet survive if you deleted the word "AI" from it?
If the gap, the magnitude, the cost, and the trade-off all still stand on their own — it's strategy. The AI is just the tool you chose to execute it.
If the only thing holding the bet upright is that AI is in it — it's a proxy. Again.
Strip the costume and one of two things happens.
The bet collapses, and you've just saved a year of margin and a hard board meeting.
Or it holds — and now you can defend it to a CFO who stopped being impressed by the word "AI" a long time ago.
That's the whole job. Not knowing what AI can do — that's table stakes now. Knowing which gap to point it at, and being able to say why.
A question to think about:
💬 The last time your team added AI to the roadmap — was it closing a named business gap, matching a competitor's announcement, or answering pressure for an "AI story"? Be honest about which of the three it actually was.
Hit reply and tell me — I love hearing your thoughts.
This is the exact work in my next Lightning Lesson.
On July 7, I'm running Kill the Wrong Bets: 6 Questions Real Strategy Must Survive — a live session where we put your actual top initiative through the six questions an executive uses to separate a real bet from a fashionable one. AI or otherwise.
You'll walk out able to strip the costume off any bet on your roadmap — and defend the ones that survive.
Save your seat here:
Until next week,
Elena Leonova
Executive product & business-strategy leader
I work with senior product leaders, Directors, and VPs to help them master product strategy when decisions are high-stakes, ambiguous, and made at scale - where trade-offs matter and the cost of getting it wrong is real.
This newsletter reflects the thinking behind my work across:
• Product Executive Coaching - From PM to Product Executive (Maven cohort)
• Advisory & coaching - product strategy and executive decision-making
• Writing & research - including my forthcoming book The Art of Platform Products
Maven: https://maven.com/elena-leonova
LinkedIn: https://www.linkedin.com/in/elenleonova
Website: https://elenleonova.com