As AI platforms reshape brand search, Catrióna Campbell warns that optimising for AI answers is futile without first defining what a brand legitimately stands for in its category.
When a consumer asks an AI for the best Irish whiskey, the best family car, or the top things to do in Dublin, the answer shows no clear relationship to advertising spend. I’ve run structured discovery tests for clients across multiple categories – hundreds of prompts across ChatGPT, Perplexity, Copilot and Claude – and the same pattern appears every time: a small number of sources account for most of the answers. None of them are controlled by brands. None of them are campaign outputs. This isn’t a quirk of one platform. It’s how the system works.
When these platforms generate an answer, they are assembling a response from what they determine to be credible, repeatable information. The work that influences those answers isn’t advertising – it isn’t even your website. It’s the expert review written three years ago. It’s the buying guide that ranks consistently. It’s the forum thread that keeps getting referenced. Brands have been investing in reach. AI is rewarding something closer to depth.
What makes this especially challenging for marketing teams and their agency partners is that the system is still evolving. No one has a complete playbook – despite the growing number of LinkedIn “experts” suggesting otherwise. Outputs vary depending on how the question is asked, when it’s asked, and how information is retrieved. And with advertising now being introduced into AI, the rules may shift again.
The scale of AI use is already significant. ChatGPT alone reaches hundreds of millions of users every week, and Irish research from Core (2026) suggests 65% of people have used AI tools. We don’t yet have a proven link between AI visibility and sales. That doesn’t make this marginal. It makes it early. People are increasingly using these platforms to form shortlists before they visit a website or walk into a store. AI is already shaping the consideration set.
The industry response is forming: AEO – optimising for AI answers. It is directionally right, but from my perspective it is incomplete. It is a tactical response to what is fundamentally a strategic question. You cannot optimise for AI answers if you have not first defined what you want those answers to say. Why should your brand be the answer to a category question? On what grounds? What are you genuinely best at, in a way that is credible beyond your own communications? Without that clarity, optimisation becomes an attempt to amplify something still vague or unresolved.
In my mind, AI visibility is a brand health indicator—possibly the most honest one we have. It does not reflect what brands say about themselves. It reflects what the world says about them. It aggregates third-party signals into a single, immediate answer. It doesn’t reward what brands claim. It reflects what they’ve earned.
And that has a consequence. If what your brand stands for is fragmented or weakly evidenced, it doesn’t get misrepresented—it simply doesn’t get selected. In testing, you don’t see confused brand narratives in AI answers. You see absence. The shortlist forms without you.
The starting point for brands is an honest assessment of where you actually are.
If your brand position is clear and well evidenced but not showing up in the places AI draws from, that is an execution problem. The signal exists – it just hasn’t travelled. There are specific tools and tactics for that, including digital PR, targeted coverage in specialist publications, and building presence in the review platforms and communities AI consistently draws from.
If the position itself is unclear or inconsistent, those tools won’t fix it. That work has to come first – defining what your brand legitimately stands for in its category, and ensuring it is evidenced beyond your own communications. This isn’t about campaigns. It’s about your genuine product advantage and real point of difference – what you can honestly own, and what others consistently say about you.
There is also a category effect worth understanding. Based on early testing, AI tends to stabilise a view of what matters in a category—the same attributes and the same “reasons why” appearing repeatedly. If your brand isn’t strongly associated with one of those, it risks being excluded entirely. You cannot position your way into a category attribute you haven’t earned. And if every brand is saying the same thing, AI will default to the ones with the strongest and most defensible proof.
Before asking, “how do we improve our AI visibility?”, we need to ask “what does AI currently say we are, is that true, and is it what we want?” That answer isn’t shaped by internal narratives or campaign intentions. It’s shaped by what is understood and reinforced externally.
Campaigns are episodic. They start and end. What determines whether you show up in AI is cumulative – built over time, through clarity and proof. The brands that take control of that will have a structural advantage in how they are discovered.
The rest won’t be misrepresented. They simply won’t be there.
Catrióna Campbell is founder of Tempered Thinking, an independent brand strategy practice. temperedthinking.com
