“No es amor, es una obsesión”
How to market an AI product? I don't have a playbook, but I keep coming back to one distinction that seems to matter more than it looks: the technology and the product it powers are not the same thing to market.
For many companies, and even the biggest ones, a key product marketing challenge is to tell the market that "they do AI", "they have AI", that their product and technologies are "powered by AI". Every marketer will find themselves at some point in time with these two letters to fit in somewhere, somehow, as a noun ("Sovereign AI"), as an adjective ("AI-powered"), as a complement ("thanks to AI"). Hairdressers face the same challenge placing "hair" everywhere they can in their brand names, even in French. In Tulle yesterday (a small town in the South of France), we walked in front of a shop called "Volt'hair" (poor Voltaire). Enough of silly comparisons, let's focus on the real question: how to market an AI product?
When AI is the tech
As a product marketer you may have to define an AI that is essentially a technology, ideally in non-technical terms, or terms that the target audience will understand. For instance: what is AI (insert definition), or its attributes: frontier, state-of-the-art, sovereign, etc, what it looks like through impersonation (your grandfather's first name, Claude; an animal, Le Chat; a thing, like Gemini). Given that AI is an abstract concept, you'll soon drift towards the attributes of the vision behind the said AI: sovereign, open source… where company attributes and their AI will conflate. The challenge is that this semantic space is essentially finite, scarce and abstract. In the end, you circle around an object that you can't really grasp, so you use workarounds or brand attributes. Which leads to the following question: are we talking about product marketing or tech marketing? When marketing AI as a technology, you're selling a set of (infinite) possible use cases. When marketing an AI product, you market a defined use case. The AI tech attributes remain voluntarily open to let prospective buyers imagine their own applications, which is the cause of distress (what are my use cases when so many are possible?) and limitless explorations (for instance with Narra). AI as a tech is a capability; AI-powered products are a promise of delivery. It's easier to market something you can do; it's harder to market something you could do.
When AI is the product
A product is a solution that solves a customer problem. When AI is powering the product, you'll find multiple complements such as "powered by", "supercharged by", "enhanced by", "turbocharged by", "made with". AI is an add-on, a technical enabler, sometimes the cherry on the cake. In this space, it all comes down to how much you're able to stick to your target audience without being fully distracted by AI. If you know your customer problem, you can be targeted in your product marketing, addressing customer pain points and how the AI product solves them, sticking to the principle that the problem should always precede the solution. The question of how you deliver this promise becomes secondary. Or at least it should.
The risk of adding AI to your product marketing is obviously to replace the purpose with the tech, the delivery outcome with the capability promise, to create AI fatigue and miss the point of addressing customer pain points. In this instance, AI can be a real distraction. Let's also not forget that product marketing and communication are two distinct disciplines.
Who actually cares about AI?
It depends on who you speak to. Investors and company boardrooms will pay special attention to their AI readiness as a market differentiator, a proof of innovation, even something reassuring for their prospects. It's a tag (and a reality) to validate that your company belongs to the latest wave of technological innovation. Customers might even buy services from an AI-first company to tick the AI budget box. It could even be that AI replaces the core definition of the product.
But end users and customers may value that, ignore it, or even reject it as what they have in mind when buying a service or product is the promise of delivery, however it's performed. Some might reject the idea that AI is used to deliver a service for political reasons, in which case AI becomes a liability. When AI is not a core tech but a product enabler, it can range from a mere distraction to a messaging add-on, an innovation tag, or a liability. "Caution: use at your own risk" could be said about AI in product marketing, as you'll potentially have to justify its use, be transparent about what it does, get challenged and trigger scrutiny.
If AI improves your product, it also exposes your brand.
Three questions I wish I asked myself earlier
Acknowledge: Are you marketing the promise of an AI tech possibility, or the promise of an AI-powered outcome? The corresponding PMM motions will differ greatly.
Segment: Yes, AI matters, in so many respects. But to whom it matters is the real question. Do you lead equally with AI in your investor presentations, sales pitch decks and internal memos?
Differentiate: AI can be a differentiator when it's a specific usage of the tech (generative vs. agentic, symbolic vs. neuro, French or Chinese, you name it), it becomes less of a differentiator when it's used across the board. If everyone does AI, how can you compete in a finite space of adjectives and qualifiers, where every SaaS B2B site looks like the exact copy of its neighbour? This is where moving from claims ("we are AI-powered") to a story becomes valuable: what's the problem, what's the enemy, how do we solve it.
Getting this right isn't about having the best AI, it's about being deliberate on which side of the tech/product line you're speaking from, and to whom. Is there a need for companies to root deeper into their identities, when AI drags them away from their DNA and what they have been designed to do in the first place: solving customer problems?
Let's now imagine Aventura singing about AI: "No es amor (lo que tú sientes), es una obsesión."