There’s a perennial tension that underpins many marketing leadership debates: The tug-of-war between long-term brand building and short-term performance marketing.
We’ve all been there … watching passionate chief marketing officers (CMOs), performance directors and agency leaders fiercely defend their positions on conference panels, each armed with data, frameworks and conviction.
I’ve always leaned toward the Binet and Field school of thought. That neat ‘steps’ chart showing how brand drives performance? It’s been a mainstay in my stakeholder conversations, quietly tucked in my back pocket for years. But recently, I felt that familiar framework beginning to fray.
HSBC recently hosted the first Financial Services Marketers Leaders Conference in Dubai. At the event, a London-based video and content agency, Casey Wishart, presented a provocative idea that large language models (LLMs) are not just tools or enablers. They’re now a channel. More radically, they are also an audience. That took me a minute to fully register.
We’re all adapting to a media landscape where younger audiences increasingly bypass traditional discovery channels. They don’t browse; they ask. Why scroll through pages of blue links when a single prompt to a chatbot can deliver context, options and a decision pathway? In that world, LLMs act as gatekeepers. They shape perceptions. They synthesise information. They may well decide which brand gets mentioned and which gets forgotten.
So then, is the question we should be asking: What influences the influencer… when the influencer is an algorithm?
This is where the debate about brand versus performance begins to get a bit knottier. Because if LLMs are to be understood as a new type of ‘media’, we have to play by a very different set of rules. You don’t buy your way into their favour through paid impressions or carefully calibrated bidding strategies. Instead, you go old school.
Editorial content, white papers, thought leadership – the unsexy stuff we scorned in the early 2000s because it didn’t generate clicks fast enough – are now some of the most influential materials in an LLM’s diet.
Casey Wishart noted that Gartner predicts brands’ organic search traffic will decrease by 50 per cent by 2028, which is right around the corner, as people embrace AI powered search. PR suddenly matters again. Authority matters. Citations, credibility, clarity of purpose are the signals LLMs are tuned to absorb. That means brand marketers might be better placed to influence LLMs than performance teams – at least in their current form. And here’s the real kicker: while we can’t yet track ‘conversions’ from an LLM response, we all know that change is coming fast.
It’s only a matter of time before those answers are embedded with shoppable actions and transactional pathways, or logged in first-party data ecosystems. There’s a whole new ‘Page 1’ battle commencing, and I’m not sure that many marketers are fully armed.
To be clear, this isn’t an argument against performance marketing. But let’s be honest: it’s in the middle of an identity crisis. The obsession with click-through rates, attribution models and return on ad spend (ROAS) dashboards is starting to look a little brittle in the face of cookie deprecation, regulatory shifts and rising consumer cynicism.
The way we operate has to evolve, bringing together performance with brand authority, speed with credibility. And as some of the big tech platforms lose relevance or battle credibility issues, performance teams are being forced to rethink what success really looks like. Increasingly, the answer is: trust.
As marketers, we’ve always believed in the power of storytelling. Now we must learn to teach as well. These models learn by crawling, reading, indexing and summarising. So the question for every brand becomes: What are we teaching them about us?
This is not a job for the data science team alone. It’s not just a compliance issue. It’s a brand issue. A trust issue. And that lies firmly within marketing’s remit.
We’re entering a future where every brand interaction might be filtered through an AI lens – whether that’s a customer asking ChatGPT about mortgage options, or a procurement team querying an enterprise LLM about environmental, social and governance (ESG) credentials. If your brand isn’t represented clearly and credibly in the data, you may simply vanish from the shortlist.
So yes, brand still matters. performance still matters. But the battleground is shifting. The LLM layer adds complexity, but it also offers a new frontier – one where core principles such as consistency, clarity, credibility and trust come roaring back into relevance.
We must be ready to meet that challenge, to reframe how we think about influence and to ensure that when someone, or something asks who we are, the answer is accurate, powerful and aligned with everything we’ve worked so hard to build.
By Aimee Peters, Regional Head of Brand, Partnerships and Wholesale Marketing, MENAT, HSBC
