When the buyer is a machine that thinks like a human, what breaks the tie?

Buyers lean on AI-generated answers more and more these days, but when a business realises it’s not featuring in those results, the instinct is almost always to go and ‘fix’ the website.

They rewrite the copy, sharpen the messaging and tidy up positioning. It is a logical reaction, although it’s usually only half the answer, for reasons that have less to do with the website than with the machine user now reading it.

Buyers use AI to take the legwork out of early research. AI is great at sifting through the market and handing back a shortlist before anyone’s sales team has even had an opportunity to connect.

It’s worth being clear about what that reader is, because it isn’t a search index politely matching keywords. The models doing the initial heavy lifting are trained on us, so they tend to size up a business the way a person would when they’re choosing between suppliers – biases, mental shortcuts and all.

A person choosing between suppliers does it in two moves, not one. First, they work out who can be trusted and then, from the names left standing, they work out who is actually the best fit.

The first move is about trust, and this is where the instinct to fix the website is right. The widely accepted GEO (generative engine optimisation) advice is to publish evidence rather than opinion: verifiable proof such as reviews, case studies, third-party ratings and analyst mentions. The machine isn’t just retrieving those facts, it is forming a view on whether a business can be believed, and the claims it can’t substantiate aren’t marked down in an evaluation matrix, they’re simply skipped as if they never existed. This is the more dangerous outcome, because a business never sees the shortlist it didn’t make, and so never understands what it was left off.

But clearing that bar only gets a business into the room. The next bit is the part the website fix doesn’t solve, because once you and three comparable competitors have all published the reviews and the case studies, evidence stops setting you apart and simply becomes the cost of being considered. Everyone still in the running has proof. What none of them can conjure on demand is a genuine reason to be picked over the others.

And this is where all the website ‘fixing’ comes undone. For example, a remarkable number of firms describe themselves as a ‘trusted adviser’, a phrase so universally adopted that it has stopped meaning anything at all. It is the beige curtain of archetypes: the safe, agreeable default that feels like a point of difference without being one. But to differentiate is to be genuinely different, and to keep competing on that difference it has to be defensible – the sort of thing a rival can’t simply copy onto their own website by Friday. When the machine weighs two providers making the same interchangeable claims, those claims cancel each other out. Sameness reads as noise. What breaks the tie is whatever is truly distinct and can be backed up.

The trouble is that plenty of businesses don’t actually know what that distinct thing is. Or they think they do, and it turns out to be an internal story that stopped matching reality a while ago. That answer is rarely found in a strategy session around a whiteboard – it lives with customers and out in the market, in the language people use about a business when it isn’t in the room to steer them.

The good news is that most companies are already sitting on the raw material without realising it. Sales calls, customer interviews, win-loss notes, support conversations, the survey that got run once and filed away, are together a running record of how a market genuinely describes a business, and why it picks one provider over another, sometimes in words the business would never have chosen for itself. The trick is to continually measure that data together, rather than commissioning it once a year and then forgetting about it. Done properly, it shows what a business is really known for, where perception has drifted from intention, and which of its claims would survive contact with a sceptical machine.

So the website was only ever half the job. AI has taken an old habit – running a brand on instinct – and made it visible in real time, right in front of the buyer. Cutting through the machine separation requires businesses to know – and be able to prove – what genuinely sets them apart, rather than hoping nobody checks.

Raine Gaisford is the founder of B2B marketing agency LimeHub and brand intelligence platform Brander.

Image: Supplied