Get ready for the AI agent revolution — and why search and display are under threat

Opinion

The ability of AI agents to streamline tasks poses a fundamental challenge to businesses reliant on web traffic and highlights the need for innovative approaches to engagement.

I have a hypothesis that I’m playing with — that AI will likely drive advertising spend towards OOH and other media including TV, video and radio. The biggest impact will be the move away from display and the evolution of search.

With the gradual removal of third-party cookies from platforms like Google Chrome, aka “the cookie apocalypse”, tracking user behaviour across the web will become increasingly challenging.

This shift threatens to drain billions of dollars away from the display ad market, prompting advertisers to explore alternative avenues for reaching consumers.

Changing human behaviour

But this is not the end of the challenges for display — the cookie apocalypse will soon be followed by “the agent apocalypse”. OK, that’s a poor name for it, but let me explain…

The way consumers interact with information and brands is going to evolve.

One aspect of AI that people have already started to become familiar with is, of course, large language models (LLMs). They’ve been around for a while now, with ChatGPT currently proving very popular — these AIs are amazing in the way they understand and can respond to you*.

But there are other rapidly developing areas of AI that are going to fundamentally change the way we engage with services and information. At the cutting edge of this are AI agents.

One port of access

AI agents supplement LLMs in that they can execute tasks on behalf of users. Instead of accessing lots of websites and apps in order to conduct a task, AI agents will act as your assistant and offer you one port of access to everything — all the information and services that you need to research and complete your task.

Companies such as Rabbit (in collaboration with Teenage Engineering) are pioneering devices like the R1 that will streamline everyday activities, from booking taxis to planning vacations. It aims to replace your mobile phone and completes your tasks at the press of a button. Importantly, it isn’t just about an LLM; it’s about a large action model that you (and others) can train to complete tasks.

Google (and potentially even Apple) is working on its own AI agents — one example you can try today is booking a holiday via Gemini.

All you need to do is give a rough brief (eg. potential locations and dates) and Gemini will do all the research in seconds, suggest itineraries and prepare everything ready for you to book. Currently, that last booking step has to be done by you, but I’m sure this will change soon. What is interesting here is how, as a consumer, my relationship with information and brands could fundamentally change.

Historically, we’ve used our laptops to go to websites to explore. We’ve read travel blogs and publications, browsed different travel companies and booking services. Just think of the hours that we have spent and the number of banners and MPUs (mid-page units) that have been presented to us.

If we can ask a device or technology to do this work, how do we know the very concept of websites will survive the AI agent revolution? We don’t. This poses a fundamental challenge to businesses reliant on web traffic and highlights the need for innovative approaches to engagement.

Jobs to be done

There’s a methodology called “jobs to be done” around product innovation and development. A classic example is lawn-mowing. A manufacturer could look to innovate by designing lawn-mowers that are automated.

But what happens if a company creates a different kind of innovation — a genetically engineered grass that grows perfectly to three inches and stops at that height? It would destroy the lawn-care industry.

We’ve been asking: how can I cut grass to make it better? But that’s not the job that needs to be done — the job that needs to be done is to create and maintain a nice lawn.

The question is to what extent the AI agent revolution will have a similar effect on websites. We previously used libraries, bookstores and travel agents to source information and book a holiday; now we use blogs and booking websites. But is that going to be the case in the future?

The “job to be done” here is to book an amazing holiday; it isn’t to spend hours using platforms trying to find the best itinerary and the perfect deal (and see all of those display ads in the process).

I expect AI agents to become ubiquitous tools for consumers. As we approach the release of groundbreaking products that consolidate services into a single interface like the R1, the advertising landscape also stands on the brink of transformation.

The future of search and display

Search is likely to remain strong, but it will change from the Google search we know today to sponsored recommendations from AI agents. Search will evolve from search advertising to AI recommendation advertising.

The cookie apocalypse, rapidly followed by the advent of AI agents and its impact on consumer behaviour, means display advertising is going to be hit hard — and it may never quite recover.

My hypothesis is that because advertising channels such as video and audio are entertainment media already tied down by copyright, AI is less likely to easily collate, share or adapt them. Meanwhile, OOH will not be greatly affected by AI agents because it is a real-world medium that is intrinsically linked to geography, point of interest and point of sale.

By championing innovation and reimagining their approach to consumer engagement, advertisers can position themselves at the forefront of this technological revolution. From ad executives to curious consumers, we must all embrace the transformative power of AI to thrive in the digital age.

* As an aside, I’ve particularly enjoyed speaking to ChatGPT about history. I like to think that asking ChatGPT about history is akin to meeting a drunk history professor (or somebody who claims to be a history professor): they can tell you incredible stories about anything, but you can’t really trust what they’re saying — at least not for certain. 

Jon Block is chief product officer at VIOOH