Shirley Marschall is back, and in this week's column she's looking at the role of agentic ad tech, AI and the need to look beyond the hype...

When AI is a bubble, and talking about AI being a bubble is a bubble… what do you do? Right, you start talking about AI agents.

And AI… agentic… what does it matter? Once you put out a new message, you quickly find a small group of people most likely to respond. You harvest that group fast, performance drops, you change the message, find a new cohort, repeat.

Talking about it would be one thing. What is happening right now is something else entirely: the line between a bull market for agents and bullying everyone into believing agents are the thing is getting blurry. And the rhetoric always has to be so bloody dramatic: an earthquake, a paradigm shift, another revolution...

Yet watching it unfold often feels a lot more like watching a soap opera-level drama.

Anyway, here we are. Ad tech has moved on from AI to agents as well. Agentic solutions are popping up all over the ecosystem, and a handful of OGs created a mini hype-cycle around it. How? By officially adding an ”agentic“ protocol to the mix, declaring it to be the next big thing, and hoping enough people believe it so the hype seeps into the rest of the world. A self-fulfilling prophecy. Hype by committee. Collaboration by comfy consensus.

So let’s get this straight. Agentic is supposed to make it easier for buyers to discover audiences and buy ads across platforms in a standardised way. Sounds a bit like launching a brand-new invention called Kettle 2.0 and proudly announcing it ”boils water“, no?

Sounds familiar? It was literally the promise of programmatic. Luckily, agentic ad tech is here now to save us from the complexity programmatic created.

And yes, it’s being sold as something different. More agentic, more context-aware, more interoperable. In plain English, MCP (Model Context Protocol) opened this door, and now everyone wants to be the ChatGPT of ad tech. It is seductive, this idea of one interface to rule the open web, and everyone wants to become the ONE.

But this is what the ad tech industry does best. It loves shiny things. It takes an acronym, glues another acronym to it, sprinkles investor money over it, and voilà, another layer of inefficiency masquerading as innovation.

Maybe asking what none of these fancy agentic ad tech solutions do is the better question? None replaces RTB. None reinvents auctions. None solves fraud, transparency, or market concentration. They just add another layer that talks to the existing layers.

Meanwhile, the industry is being told it needs this layer to unlock programmatic efficiency, signals, and whatnot. Yes, this agentic one. Which sounds clever until you realise what it actually means: let’s put another middleman between you and your money. Again…(or when ad tech became ad tax.)

To be fair, the idea of agents talking to each other sounds interesting, even futuristic. But it also sounds very technical and intimidating for non-techies. The bottom line question is: what does it actually unlock compared to what we already have? Is all this AI / agentic frenzy improving the fundamentals of marketing? It should be making things simpler, not more complex. Right?

Cameron Armstrong made a brilliant comparison about ad tech that fits the agentic bill perfectly: ”You order a pizza and only 44% arrives. You’d send it back, right? Yet in advertising, we keep paying the bill. Every few months, a new report reminds us: budgets are leaking, transparency is missing, and trust is thin. But what do we do? We nod, we share the article on LinkedIn, and then we go right back to ordering another half-baked pizza.“

So how about we don’t make the same mistake with AI and agentic. Because here’s another hard lesson learned, quoting Mark Pilipczuk: “Whenever we hear hype, the real question is cui bono? And if the ad tech bros are hyping it, the cui is ad tech bros and the bono is big.“

Yes, the architecture of innovation still comes with a toll booth.

And yes, it’s entirely possible this critique is missing the point, or the promise, or both, or that nobody quite knows what those are yet. And that’s exactly the bull vs. bully issue. As Erez Levin put it, the industry would be far more receptive if there were more humility. If it felt like everyone was being invited to understand the visions and the backend, not just the shiny frontend, and to join the movement rather than being told “this is the future, 100%, get on board or get left behind.“

Maybe what this industry needs most right now are some agentic playground rules. No panic. No 100% promises. No underestimation. No drama. And a lot less hype…