Your AI strategy is a press release.

Last fall I sat in a quarterly review where the CMO pulled up a slide titled "AI-Powered Growth." It listed seven initiatives. Personalized recommendations. Dynamic creative optimization. Predictive churn modeling. Conversational commerce. The works. Every initiative had a vendor attached and a launch date penciled in. What it did not have was a single reference to a customer problem any of these things were supposed to solve.

I asked the room a simple question: which of these seven things did your customers ask for? The silence told me everything.

This is not an unusual scene. I've watched it play out at consumer tech companies, outdoor brands, mid-market retailers, and hospitality groups over the past eighteen months. The pattern is always the same. Someone in leadership gets the memo that every serious company needs an AI strategy. A task force forms. Vendors pitch. A plan materializes. And then the company announces it, often before a single customer has been affected by any of it.

The announcement is not the strategy

There is an important difference between having AI capabilities on your roadmap and having an AI strategy. A roadmap is a list of things you intend to build or buy. A strategy is a theory of how those things create value for a specific person in a specific context. Most of what I see companies calling their AI strategy is actually a procurement plan with a narrative wrapper.

The tell is in the language. When the strategy deck leads with "we are integrating AI across our platform" rather than "our customers struggle with X, and we believe AI can address that in this specific way," you are looking at a press release draft, not a strategic document. The technology is the subject of the sentence when the customer should be.

I get why this happens. The pressure is real. Boards are asking about AI. Investors want to hear the word in earnings calls. Competitors are making announcements, and silence feels like falling behind. But responding to that pressure with an announcement is not the same as responding to it with a plan. It just feels like it in the moment.

The checkbox problem

I worked with an outdoor brand last year that had added a chatbot to their product pages. When I asked how it was performing, the team showed me completion rates and containment metrics. Fine. When I asked what customer need it was addressing, the answer was vague. Something about reducing support tickets. When I pulled the actual support data, the most common questions were about sizing and fit for specific activities. The chatbot wasn't trained on any of that. It was answering questions about return policies that were already clearly stated on the site.

The chatbot existed because someone decided the brand needed a visible AI touchpoint. It was a checkbox. It let the team say "we have AI in the customer experience" at the next all-hands. But it wasn't solving a problem. It was occupying a space.

This is what I mean by the checkbox problem. AI becomes a line item that needs to exist rather than a tool that needs to work. The metric of success shifts from customer impact to organizational optics. Did we ship it? Yes. Is it AI? Technically. Can we talk about it externally? Absolutely. Is any customer's life materially better because of it? That question doesn't come up.

Capability is not strategy

The technology vendors have made this worse, though I don't blame them entirely. When every platform you already pay for is adding AI features, it's easy to conflate adoption with strategy. Your CRM has AI now. Your analytics suite has AI. Your creative tools have AI. So by extension, you have AI. Right?

Not really. Having access to AI capabilities is like having access to a gym. It doesn't mean you're fit. Fitness requires a plan that accounts for your specific body, your goals, and the discipline to follow through. Turning on the AI features in your existing stack is not a strategy. It's a software update.

I've seen a consumer tech company proudly announce AI-driven personalization that, in practice, was recommending products people had already purchased. The algorithm was running. The feature was live. The experience was worse than the static merchandising it replaced. Nobody caught it for weeks because the team was measuring whether AI was on, not whether it was good.

What a real AI strategy looks like

A real AI strategy starts where all good strategy starts. With the customer. Not with the technology, not with the competitive landscape, not with the board's expectations. With a clear-eyed understanding of what your customers are trying to do and where they are struggling.

From there, you ask a harder question: is AI actually the best way to address this? Sometimes it is. Sometimes a better FAQ page, a clearer navigation structure, or a well-trained human support team would do more. The discipline is in being honest about when the technology fits and when you're reaching for it because it's fashionable.

When AI is the right answer, a real strategy gets specific. It defines the use case narrowly. It identifies what success looks like from the customer's perspective, not just the business's. It builds in feedback loops so you know whether the thing is working or just running. And it treats the first version as a hypothesis, not a finished product.

I worked with a hospitality group that did this well. They noticed that their loyalty members were consistently frustrated by the gap between their stated room preferences and what they actually got at check-in. Instead of announcing a broad AI initiative, they built a narrow model that cross-referenced preference data with real-time inventory at specific properties. It didn't get a press release. It got a measurable improvement in guest satisfaction scores at four pilot locations. Then they expanded it. That's strategy.

The cost of getting this backward

The risk of treating your AI strategy as an announcement isn't just wasted budget, though there is plenty of that. The deeper risk is credibility. Customers are not stupid. When they interact with an AI feature that clearly doesn't understand them, that recommends things that make no sense, that adds friction instead of removing it, they don't think "well, the technology is early." They think the brand doesn't understand them. That's a positioning problem, not a technology problem.

Internally, the cost is cynicism. Teams that ship AI features they know aren't solving real problems learn to stop asking whether something is worth doing. They learn to ask whether it's on the roadmap. That's a culture shift that's hard to reverse, and it makes the next genuinely good idea harder to execute because nobody trusts the process anymore.

I'm not arguing against using AI. I'm arguing against announcing it before you've earned the right to. The companies that will look smart in three years are not the ones that said the most about AI in 2024. They're the ones that quietly figured out where it actually matters for their customers and built from there.

A press release is easy. A strategy is earned.