• by Maarten Albarda , Featured Contributor, 3 hours ago

We promised to follow my look back on 2025 last week with some ideas of what will be important in 2026.

It probably won’t surprise you to learn that, in 2026, everything will be AI and AI will be everything (everywhere, all at once).

This week, The Wall Street Journal published a fairly gloomy op-ed titled "AI Is About to Empty Madison Avenue.” It was written by Columbia Business School professor Rajeev Kohli. He argued that because platforms like Google’s Performance Max and Meta’s Advantage+ are automating the creative process, the agency model is doomed. He’s predicting a ghost town of displaced copywriters, art directors, and other creatives.

If you look at where the smart money and the serious conversations are going, we are moving past the "wow" phase of generative AI and entering the era of agentic AI. This was clear from the presentations at the recent Reuters Momentum AI conference held in New York. The buzz wasn't about chatbots; it was about autonomous agents -- software that doesn't just "chat," but actually does things. There was a great deal of talk about this so-called dual transformation, where firms are reinventing core operations while building AI-native business models.

This is why I disagree with some of Professor Kohli’s conclusions or predictions. The real disruption isn't that a computer can write a tagline. It’s that a computer will orchestrate the entire messy, disjointed workflow of modern advertising.

Take the startup Fluency, for example. The company just raised $40 million to double down on this exact concept. It isn’t trying to be another tool in the ad tech stack; it’s positioning itself as a Digital Advertising Operating System. In other words, not part of the tech stack, but THE tech stack. Fluency's goal is to centralize campaign execution across the walled gardens (Google, Meta, Amazon) and the open web.

This is obviously the Holy Grail many are chasing -- WPP, Stagwell and Publicis, to name just a few. Marketers and agencies currently are forced to have team members on Meta using Advantage+, others on Google using PMax, and yet others managing a DSP for the open web. It’s a logistical nightmare. If agentic AI can actually unify these into a single, autonomous workflow, we aren't "emptying" Madison Avenue of creativity; we are emptying it of fragmentation and ineffectiveness.

The word “toil” came up repeatedly during the Reuters event. The consensus among the 300+ leaders at that event was that AI creates a "human-plus-machine" operating model. It automates the day-to-day and mundane so humans can get back to the strategic, relationship-driven work that actually moves the needle. At least in the short term -- all this could just be wishful thinking.

So perhaps we should stop treating AI for now as a content factory and start treating it as an operations manager. Look at your workflow friction points. Where does data get stuck between platforms? As some at the Reuters event noted, "garbage in, garbage out" is the biggest barrier to scaling AI tools.

So for 2026, ignore the most gloomy of predictions about the impact of AI. The agencies of the future will be smaller, yes, but they will be significantly faster -- and, hopefully, a lot more interesting to work for.