In part one I introduced the Agentic AI Asymmetry Dilemma, the structural gap between agencies building proprietary AI and the publishers trying to optimise for systems they can’t see. This piece turns to advertisers. Because the asymmetry isn’t just a publisher problem. It runs straight through the advertiser relationship too.
The Numbers Are Already Moving Without You
The scale of what’s happening is worth pausing on. According to Digiday’s State of Agentic Advertising report, more than 90 per cent of advertisers now use AI to plan media, set budgets, and optimise targeting. McKinsey’s February 2026 survey found 34 per cent of enterprise marketing teams are already running at least one autonomous agent in production, more than double the figure six months earlier.
The question needs to be asked are advertisers driving this AI advancement, or are their agencies?
AI search traffic is predicted to surpass traditional search by 2028. Adobe found AI traffic to US retailers rose 393 per cent in Q1 2026 and converted 42 per cent better than regular traffic, a complete reversal from twelve months prior.
This is not a trend you can watch from a distance. The systems making decisions about where your budget goes are already running.
The Advertiser Side Of The Asymmetry
Agencies are building proprietary AI systems that plan media, allocate budgets, evaluate publishers, and automate significant parts of the advertiser-agency workflow. In my experience most advertisers have limited visibility into how those systems are built or what criteria they use. Worse, many have structural blockages inside their own organisations that prevent them from taking advantage of these capabilities even when they want to.
Bain & Company’s recent article, Proprietary Intelligence: How to Win with AI, describes the advantage the best-performing AI organisations are building as “proprietary intelligence”.
This is the combination of unique data, encoded workflows (the institutional knowledge of how the organisation actually operates, embedded into agents that act on it at scale), and a learning architecture where every deployment gets smarter with use. That is precisely what agencies are building. And it is precisely what most advertisers are not.
The result is a familiar but deepening imbalance. Advertisers are over-investing in implementational work; repetitive, duplicative tasks that absorb agency time and advertiser budget without generating strategic value on either side. The agency spends less time solving business problems. The advertiser spends more time managing process.
The fix is not a cost review. It is a workflow conversation. When advertisers and agencies build a shared view of how their combined workflow operates, and how AI will support and automate that workflow, both sides benefit. The agency gets time back for the thinking it does best. The advertisers reinvest what was spent on implementation into the strategic work that actually moves the business forward.
The Wrong Question
Cost pressures and short-termism are pushing the whole industry toward an unsustainable place. Advertisers are cutting costs. Agencies are losing margin. Publishers are getting squeezed. The Basis 2026 Agency Report found 87 per cent of agency professionals believe the traditional model is broken or racing toward it.
We are collectively optimising the current system into failure rather than building the next one.
Advertisers who cut costs and expect their agency’s AI to compensate aren’t solving an execution problem; they’re outsourcing the work and calling it a strategy.
AI can absolutely help reduce the cost of low-value transactional work. But the savings should be reinvested into the relationship and more sophisticated technology, tools and capabilities, not extracted from it.
Crap In, Crap Out. At Scale
Gartner predicts more than 40 per cent of agentic AI projects will be cancelled by end of 2027 due to unclear value, rising costs, and weak governance. The sophistication of the model does not fix the quality of the inputs. It amplifies them.
Advertisers who hand over their AI strategy to their agency, without aligning the KPIs being optimised toward, without ensuring their own data is feeding the model, are not benefiting from their agency’s AI investment. They are subsidising it.
Adobe’s 2026 AI and Digital Trends report found only 51 per cent of organisations have the cloud infrastructure required for agentic AI, compared to 89 per cent for generative AI. The infrastructure gap on the advertiser side is real, and it is the gap their agency’s AI is working around, not working with.
The Blockbuster Moment
This is not the first time an industry has faced this choice.
Blockbuster didn’t fail because Netflix had better movies. It failed because Netflix understood what the technology made possible and built around it. Kodak invented the digital camera in 1975 and buried it, not because they lacked the technology, but because they wouldn’t cannibalise a business that was still working.
The technology itself is not the competitive advantage. Everyone eventually gets access to the technology. The advantage goes to the organisations that understand it deeply enough to shape it; that build the capability, the workflows, and the cultural fluency to use it as a lever rather than a threat.
AI on its own is not a competitive advantage. But no AI capability at all. No understanding of how the systems working on your behalf are built. No ability to shape the criteria they use. And no data feeding the model. That is not a gap, it is a structural disadvantage that compounds every quarter it goes unaddressed.
What Advertisers Actually Need
The Adobe 2026 AI and Digital Trends report also found that 63 per cent of organisations expect agentic AI to give employees more time for strategic and creative work. The direction of travel is collaboration, not competition; realignment of budgets, not budget cuts.
The Advertisers who will win are not the ones who outsource their AI to their agency and hope for the best. They are not the ones who resist because the current model is still “working”. They are the ones who build just enough capability to be an intelligent co-pilot, who understand the system well enough to shape it, interrogate it, and redirect it when it drifts from what the business actually needs.
The brave advertisers; those who concentrate on business metrics, develop customised workflows with their partners, and explore how AI augments rather than replaces strategic thinking, will be the ones who grow. Not because AI gave them an advantage. Because they understood it well enough to use it as one.
You don’t need to build the engine. But you do need to learn to drive.
In the final piece I’ll bring all three parties together, publishers, advertisers, and agencies, and argue that the Agentic AI Asymmetry is ultimately an ecosystem problem. One that no single party can solve alone, and one that the industry has both the incentive and the capability to fix. If it chooses to.
Stuart Bailey is the chief media officer of Clemenger BBDO.
