One practical marketing idea a week
Science-backed, practical, no fluff. Unsubscribe any time.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Prefer to talk? Get in touch
Getting cited by AI isn't a new technical game - it's the oldest game in marketing, now with a machine as the first gatekeeper. What E-E-A-T, the messy middle and mental availability have in common.
A note for marketers panicking about being “cited by AI.”
A growing number of your customers now open ChatGPT, or skim Google’s AI overview, before they ever open your website. They ask a machine to recommend, compare and shortlist - and increasingly they act on what it says. Which has sent a lot of marketers into a familiar scramble: a new channel, a new acronym (GEO - generative engine optimisation), a new thing to “win.”
Here’s the part the scramble misses. The thing that gets you cited by an AI is, more or less, the thing that’s always got you chosen by a human: trust.
Large language models don’t surface the loudest brand or the cleverest keyword. They surface the source that looks most credible, most referenced, most consistently talked about - the one with authority. Google has spent years formalising exactly this in its own results under the banner of E-E-A-T: experience, expertise, authoritativeness and trust. The AI models trained on that same web lean the same way. So if you want to be cited by AI, be citable: have a real point of view, get referenced by others, say something worth repeating.
Now look at the human on the other side of the screen. Google’s own research into the “messy middle” of the buying journey (Rennie and Murray, 2020) found that people don’t march neatly down a funnel. They loop between exploring and evaluating, sometimes for weeks, until something tips them. What tips them is rarely more information - it’s mental shortcuts.
The study isolated six: category heuristics, the power of now, social proof, scarcity bias, authority bias, and the power of free. Look closely and most of them are trust shortcuts wearing different clothes - we trust the expert (authority), and we trust the crowd (social proof).
And we trust the brand we’d already heard of before we needed it - what the Ehrenberg-Bass Institute calls mental availability. The brand that’s already in mind when the trigger hits has won half the battle before the search even starts.
So the machine and the human are hunting for the same signal from opposite sides of the glass. The AI is working out which sources to trust on the buyer’s behalf. The buyer is working out who to trust with their problem. Build genuine authority and you win both at once. Chase the algorithm with thin, over-optimised content and you lose both - because the same emptiness that fails to earn a citation fails to earn a customer.
That reframes the job, and it should be a relief. Getting cited by AI isn’t a separate technical game bolted onto marketing. It’s the oldest game in marketing, now with a machine as the first gatekeeper. The work is the work it always was: be distinctive, be genuinely useful, get referenced by credible places, be remembered before the moment of need. Publish things only you could write. Apply the messy-middle biases - but ethically, as nudges that reduce friction, not tricks that manufacture it. (Edelman’s annual Trust Barometer is a useful, sobering reminder of how little slack buyers give a brand that overreaches.)
The brands that get surfaced, cited and chosen in an AI-shaped world won’t be the ones who gamed a new acronym fastest. They’ll be the ones who were already worth trusting - and made sure the evidence of it was visible to humans and machines alike.
The channel changed. The job didn’t.