AI can fix advertising’s relevance problem – if we stop feeding into the hype
Omnicom Media UK AI lead, Sean Betts, on how agencies can seize AI’s transformational potential and finally solve advertising’s relevance problem
“I’m always very sceptical when industry people start talking about super agents and agent swarms because it all just sounds like marketing hype.”
This is what Sean Betts, Chief AI and Innovation Officer at Omnicom Media UK, jokingly said a few minutes into our chat on AI. For Betts, a leader who combines board-level AI strategy with hands-on development and testing - leading AI transformation across OM UK while actively building his own AI tools and systems - the technology is only just beginning to meaningfully reshape the marketing landscape, despite the hype-fuelled discourse suggesting that it already has.
“In practice, AI is not transforming agencies right now, it’s just augmenting,” Betts says. Like a “sparring partner,” it is helping teams summarise research, complete tasks through agents and experiment with more advanced intelligence layers. But for Betts, the keyword is still ‘experiment’. “There’s still a lot more development and build-out needed for true transformation to become a reality,” he adds.
No agency is “fully” embedding AI into current workflows yet and Betts believes the main reason is because too much time and emphasis has been placed on the technology and not on the process and operations behind using it.
“If you can’t standardise a process, if you can’t codify a task, how can you expect an AI agent to do it for you? You have to be able to define the steps and what happens in each of those steps. AI is already very good at a wide variety of individual tasks in isolation but trying to string those together, using AI to move through workflows and trigger processes - we’re not there as an industry yet.”
Embedding AI will take time. What Betts believes agencies are struggling with more, however, is how they pitch its value to clients. “The issue, for me,” he says, “is that clients are expecting financial benefits and we’re pitching those benefits over a time period we don’t yet know we can deliver against.” Client pressure is forcing agencies to make commercial commitments, he feels. “We’re essentially making bets on a technology that we haven’t seen fully delivered. The theory is there, but in practice, no one has really seen it yet.”
So, how should agency leaders begin to reframe the way they think about AI transformation, and how can they begin to implement change?
Fixing advertising’s relevance problem with AI
Globally, ChatGPT processes over 2.5 billion prompts per day, as opposed to a comparatively measly one billion in 2025. The breadth and variety of how people use LLMs like ChatGPT means one thing for Betts - the platforms are a bona fide marketing channel now.
“The industry needs to further its understanding of how they work, how brands can show up in them, how the industry can influence that, and how the business impact can be measured.”
A whole new marketing discipline can be born into the industry, Betts feels, simply because these generative AI platforms work fundamentally differently from the traditional digital algorithms and platforms the marketing world has experienced in the past.
“In the same way we built social, search and programmatic disciplines, these platforms will demand real specialism - marketers and agency practitioners who understand them and can help clients navigate them. The challenge, which is both interesting and exciting, will be enacting a much-needed industry-wide structural change. Different disciplines, on both client and agency side, will need to come together to address this - from technical people and brand teams to PR, search and media teams and beyond.”
The reason this understanding of generative AI platforms is so important is because they provide a huge opportunity for brands and agencies to solve advertising’s ever-present relevance problem.
“Whether it is through audience-based targeting, contextual targeting or other signals we try to leverage, we’ve always tried to find proxies to make sure the advertising we put in front of someone is as relevant as possible at the exact right time.”
Now, however, the industry has the technology to understand both the intent behind what a customer is asking and the context around that individual, built up over time through repeated use. For Betts, using these AI models as the middleware between customer and brand to ensure messaging is genuinely relevant is a “no-brainer.”
There are still transparency and ethical challenges to navigate, such as disclosure and distinguishing ads from organic responses, but they are solvable. “And if you solve them,” Betts says, “the industry has a real opportunity to possess the perfect mechanism to make advertising relevant again - in a way we haven’t been able to for quite some time.”
What really concerns Betts is that agencies and brands are already starting to treat these popular LLMs like “any other” digital platform, dropping display ads into them and applying basic targeting. “We have this incredibly powerful technology in the middle that we could use to make advertising far more relevant, and we’re not using it,” Betts attests. “There’s a danger we go down a familiar, well-trodden path that doesn’t work - instead of leveraging what’s right in front of us to do something much better.”
The industry now has the means to crack the relevance problem because brands can meet user intent directly and provide more useful, effective and even entertaining advertising. “With the help of these AI platforms in creating ecosystems, formats and mechanisms, brands will be able to guide decision-making, explain their products and make sure product details are up to date and available to be surfaced.”
For all of this to happen, the industry will not only need to increase its understanding of how AI can integrate into operations, but also how it can be used to create new capabilities. “For me, that’s one of the big missing parts of the narrative right now: how do you use this to build better capability, rather than just drive efficiency through a business? These AI platforms are bona fide marketing channels because they’re driving consumer behavioural change. We have to think about this as something that impacts consumers, as well as something that impacts business operations.”
This does not mean traditional channels will disappear. “People thought radio would die when TV was invented - it didn’t. Search will serve a slightly different purpose and also benefit from this surge of AI platform use, because people are spending more time online, and people who spend more time online tend to search more.”
What will change for Betts, is the nature of online traffic as less and less people visit websites. “That doesn’t mean search disappears, but it does have a potential impact on display advertising. If people visit fewer websites, they’re exposed to fewer display ads. That’s the one area that could really change. Everything else will continue - the volumes might just decrease over time.”
The industry is facing a transformation problem, not a tech one
What now for agency leaders then? The main mistake Betts sees is treating AI disruption as a technology problem rather than a transformation one. “And for our industry, that transformation is both external and internal,” Betts adds. “Externally, it’s about consumer behaviour shifts. We help clients reach consumers, so if consumers are behaving differently, we need to understand that, advise on it and help brands show up in the new places where people are spending time.”
Internally, the limitation is not the technology but rather that agencies have not codified its workflows and tasks well enough to apply it properly. “We need to make our organisations more operationally fit so we can then apply the technology effectively.”
One thing that keeps the OM UK AI lead up at night is the thought of leaders assuming they will be able to log into a platform one day and have it do everything for them. Effort, development and research, Betts says, needs to be put into building scaffolding around the technology. “When you get there, it can be transformative,” he admits. “But it will take hard work. You can’t just flick a switch and expect it to plan and buy media campaigns for you.”
Betts warns adland leaders of the obvious truth. “Even if AI ends up being able to do everything for us - the second that happens, the industry is over. So there’s no point waiting for that. You have to shape it in a way that works for you - bring your own thinking, your own intellectual property. If you put the effort in, you’ll get the reward out.”
Upskilling needs to happen from the ground up
Last year, all 1,800 OMG UK employees had to complete an ‘AI readiness’ training programme. The first phase alone featured seven hours of the training on skills, governance and ethics. This AI literacy is something the agency and others alike have placed a lot of importance on, but Betts is adamant that having practical, hands-on experience is just as important.
“I can’t describe generative AI in any other way than as a general-use technology. It’s like the internet - you can’t really teach someone how to use it. People have to go online, get familiar with it, become comfortable with it, and then work out how it can help them in their specific context.”
Betts is clear that organisations need to get hands-on at a grassroots level, figuring out how to use it in their roles, tasks, teams, markets and with their clients. “This isn’t something where you can define use cases top-down and say, “This is what generative AI is for.” That would be like trying to define what the internet was for in the late 90s - you’d look like a complete fool.”
Crucially, this is how the industry starts to solve the relevance problem in practice, not through top-down theory, but through people working out, in real contexts, how AI can make what they produce more useful, timely and meaningful.
“Over time with AI, you can start to connect use cases, identify patterns, and make it easier for everyone to adopt,” Betts adds. “But you have to start bottom-up.”
What will have changed in the 2030s?
“Probably not a lot - and a huge amount - both at the same time,” Betts jokes. Not a lot will change in traditional channels - they will experience gradual but not dramatic change, Betts feels. A huge amount of change will take place within the digital ecosystem however, with AI platforms playing a much more important role.
“The scale of usage and the amount of time people spend on these platforms means brands need to show up there - and we need to help clients navigate that in a different way.”
At the same time, Betts also believes that day-to-day work will look very different. “I don’t think it will take longer than five years for more advanced agentic AI and workflow automation to become a reality. A lot of what we’ve been talking about will likely materialise in the next two to three years.”
Ultimately, what Betts implies is that the industry’s long-standing challenge will not change: making advertising feel relevant to people. What will change is the industry’s ability to finally solve it.
Rather than replacing human insight and creativity, AI gives agencies a way to understand intent, context and timing at a far deeper level - and to act on it.
If the industry can resist repeating old habits and instead build around this new capability, Betts believes: “The opportunity isn’t just transformation, it’s relevance, finally delivered at scale.”
Built with Shorthand