AI is shaking up product marketing. Everyone’s racing to optimise for the bots, but it’s the humans you still need to convince, writes strategy director Michael Green.
Reporting from sunny Cannes, The Economist recently ran a great piece about how AI is turning the advertising world upside down. I don’t do Cannes, but back here in beautiful Bristol, one detail stuck with me: brands are now creating content not just for people, but for large language models (LLMs) — the bots crawling the internet and powering tools like ChatGPT and Google’s Gemini.
It makes sense. If these models are the new gateway to product discovery, it’s logical to optimise for them. Detailed product descriptions, structured data, lots of plain text — all good for getting surfaced and recommended. And some early-stage startups are already running thousands of LLM queries to analyse how they show up in search results, not for humans, but for the models themselves.
But there’s a catch. Even if a bot recommends you, it’s still a human who decides whether to trust you.
And this is where I think the conversation risks going off balance.
Optimising for the machine ≠ winning the sale
If the only people you’re writing for are LLMs, you end up with dry, robotic content. Sure, it’s detailed. Sure, it ranks. But it doesn’t build belief. It doesn’t spark interest. It doesn’t help someone feel like your product is the right fit or your company is worth talking to.
In product marketing, we see this all the time. A technically brilliant solution gets surfaced through search or demoed well enough on a call, but the website doesn’t connect. The story doesn’t land. The buyer leaves unsure or unconvinced. You’ve won the algorithm and lost the opportunity.
And with AI-driven “agents” already starting to make decisions on behalf of users, from shopping tools to price alerts, it’s tempting to imagine a future where the human buyer disappears entirely. But they won’t. Especially not in high-consideration categories where trust, clarity, and emotional reassurance still matter.
Two audiences, one joined-up strategy
The smart approach isn’t to ignore LLMs. It’s to recognise that they’re the top of the funnel now, not the whole funnel.
Yes, give them the detail. Translate your infographics into text. Write those longform product specs that make an LLM purr.
But don’t forget the person who clicks through, the one who actually signs the deal, swipes the card, or champions your product internally. That’s where your story needs to shift from facts to feeling. From visibility to conviction.
And that’s where product marketing comes in. Not as a bolt-on, but as the bridge between discoverability and belief.
Keep the bots in mind, but speak to the human
There’s a clear role here for marketing teams to play: shaping content that works both ways. Optimised for structure, but not stripped of soul. Grounded in clarity, but still distinctive.
Because yes, we’re entering a new era of AI-driven discovery. But your buyer still has a brain, and a budget, and a boss to convince.
And if you forget about them, you might end up with the best-ranked product no one remembers.