Stop with the hype: AI agents should feel ordinary, not revolutionary
Opinion
The real opportunity for AdTech is not louder promises, but AI tools that quietly make everyday planning and buying easier. Converge’s CEO explains.
AdTech has a habit of overcomplicating simple ideas, and, despite its wonderful possibilities, AI is in danger of exacerbating this and becoming the go-to ‘slop multiplier’ for the AdTech business.
More half-baked slogans of programmatic revolution on the ‘graffiti wall’ of AdTech is the very last thing media agencies and their bill-paying clients want.
They simply don’t have time for “new” solutions sold as “breakthrough” that come with layers of acronyms, dashboards, and inflated expectations.
Stretched teams of planners, traders and account teams need fewer pain points in their everyday work, not more. In turn, their advertiser clients want to adopt AI in ways that can actually accelerate performance rather than add complexity.
That is why much of the industry’s AI conversation has struggled to land at the agency level.
The problem is not fear of automation, but practicality. Few agency teams want another complex layer in an already crowded tech stack, or tools that need specialist knowledge before delivering value.
The irony is that the most interesting development in AI right now, agentic AI, works by doing the opposite. Its value lies in simplicity, not how advanced it sounds.
Spend time with an agency team, and you can explain what AI agents do in minutes. The real opportunity for AdTech is not louder promises, but tools that quietly make everyday planning and buying easier.
AI agents do a job, do it well, and do it fast
AI agents are not thinking machines masterminding the future of advertising; they are teams of models trained to perform specific, repeatable tasks at speed and at scale.
One might assess whether a page is suitable for a given brand. Another might predict how likely an ad is to be seen. A third might test variations in bidding strategy to learn which delivers the best outcomes.
Each is trained for a single job, but together they form a system that can automatically respond to thousands of variables in real time.
In effect, advertisers gain hundreds of tiny, task-focused assistants working constantly and invisibly in the background.
They test different placements, adjust bids, flag issues, and feed it all back into their training data to provide a continuous stream of fresh learning material.
Agents don’t work to a schedule; they do this every time an impression is up for auction, within milliseconds.
What matters is not the technical nitty-gritty of how an agent-driven system works but that it gets out of the way. When agents are doing the heavy lifting across optimisation, testing, and reporting, agency teams can step away from their dashboards and knuckle down on novel strategies to secure growth for their clients.
They can stray from the path of least resistance and diversify their budgets beyond walled gardens, confident that their tireless, agentic assistants will steer their campaigns toward optimal results.
This is where agentic AI frankly embarrasses much of what is currently being packaged as AI in digital advertising. Slapping a chatbot onto an existing platform does not change how media is bought; you’re just swapping clicking buttons for typing prompts. True agentic systems are embedded directly in the decision loop, bypassing intermediaries and gatekeepers that have only bogged down and bloated the programmatic supply chain.
Despite all the machines at work, what emerges is a refreshingly human process.
Instead of buyers having to learn the particularities and peculiarities of a platform’s logic, they can simply approach a task as if they had infinite time to do it themselves: look at the media, judge the context, consider the audience, and decide whether the price is worth paying.
Agents replicate that same thought process, trained on data and applied at a scale no team could ever manage manually.
Technology for the many, not the few
Large holding groups may have the deep pockets to afford towering tech stacks and the teams of specialists that keep them upright, but independent agencies and in-house brand teams rarely have that luxury. Stuck between a rock and a hard place of client demands and practicality, they often default to the channels and platforms that are easiest to use and offer the clearest path from exposure to outcomes.
Perhaps for the first time in adtech history, agentic AI is a technology that reduces rather than increases complexity.
Instead of spinning separate plates for brand safety, targeting, viewability, measurement, and reporting, even the leanest of lean teams can rely on a coordinated set of agents to handle these considerations within a single buying logic stem. The need for specialist internal engineers or budget-draining middleware evaporates.
What remains is a more meritocratic market. If access to advanced buying capabilities is no longer determined by scale or headcount, smaller agencies can compete on the quality of their thinking and creativity rather than the number of rows on their balance sheet.
With just a single tech partner, they can deploy agents tuned to a specific client or campaign, marrying the bespoke service indies are known for with bespoke buying technology.
It also frees agencies, big or small, from being merely ad spend money handlers. Too many agency teams have been relegated to mere resellers, shuffling spend from client to platform with little directorial control.
AI agents give (if you’ll excuse the pun) agencies their agency back, allowing them to base buying decisions on outcomes that matter to their clients rather than platform accessibility.
Agents will make adtech invisible
In other words, through the use of agentic AI automation, programmatic will no longer be seen as just a form of media, but rather as a mindset for integrating workflows and data.
The past two decades of adtech’s spiral into acronym-dense, spend-syphoning delirium have worn buyers’ patience for complexity down to a thread. Meanwhile, much of the industry has responded to customers’ calls for simplification by acquiring one another and expanding their reach across the supply chain, hoping to be the biggest fish in the pond by hoarding data and technology.
It’s an outdated mindset from a sector short on new blood and fresh ideas.
The standard bearers of the next era of adtech won’t be those who have the most elaborate systems, but those who make powerful technology feel ordinary, who have the humility to accept that all anyone wants from adtech is for it to be almost invisible. And those who understand that the future lies in building trust and giving advertisers the control to shape and scale their own digital advertising.
Turning agentic AI into an everyday planning and buying tool will require the hype-merchants to bite their tongues.
It means resisting the urge to dazzle agencies with technical jargon and instead speaking the plain language of business: reach, outcomes, and return on investment.
Most of all, using agentic AI to identify that magic moment when a connection is made, and a new buying action is initiated. Humbling intermediaries along the way and seeding systems that adapt to how agencies work, rather than forcing agencies to adapt to how systems work.
Yet I see many talking themselves into a familiar corner. By positioning AI as a sea change that demands an operational overhaul and a total backend refresh, the industry has made a wonderfully simple technology feel inaccessible.
The beauty of agentic AI is that it requires little explanation, because it already aligns with how people do their jobs. If anyone’s going to be replaced by machines, it’ll be those in the middle who try to cling to their positions and power, as AI bridges buyers and sellers without them.
Ian Maxwell is the CEO of Converge