For decades, SEO was the lifeblood of digital publishing and advertising, ensuring consumers found and visited websites and were exposed to the ads there. 

But search as we know it is being dismantled by AI. You only need to look to Google to see it. At the top of your web searches, you can now expect to see a Gemini AI summary that brings up the answers for you: consumers are finding information and discovering products through these summaries, chatbots and digital agents, without needing to visit a single link.

AI-driven search results are replacing search with a zero-click model. Bain & Company found that about 80% of search users rely on AI summaries at least 40% of the time, reducing organic website traffic by around 15-25%. Similar reports place the numbers even higher. 

For consumers, it’s a far easier search experience. No trawling through sites for answers, battling cookie opt-ins and pop-up adverts that cover the content. But for publishers and advertisers, traffic is down and ad impressions are shrinking. With no clicks, publisher ad space is worth less, and the financial model that supports them starts to break down.

This is the agentic web, the next great frontier in search and discovery.

Whether we like it or not, AI has rewritten the playbook and become the new audience. But in the wake of it, a new buzzword has cropped up: Generative Engine Optimisation, or GEO. Where SEO was about winning the clicks, GEO is all about earning the citations that come with AI-generated answers. Being trusted by the LLMs so that you appear in generated recommendations.

It’s one of the largest shake-ups in digital advertising to date. AI isn’t going anywhere fast, so getting to grips with what comes next is going to be key to staying on top.

The evolution of search

SEO, or Search Engine Optimisation, was designed for simpler times. The term came into the public consciousness in mid-1997, and became popular over the next three years as more companies became interested in optimising their websites to get more attention and inbound traffic. It relies on a simple algorithm of feeding in keywords and building backlinks to increase visibility. The user journey was predictable, with people searching and choosing which site to visit based on headlines and page descriptions.

AI search flips the logic. With generative AI, the user might not choose a site at all, instead reading a quick and clean summary of the content. If you get lucky, your content might fuel the answer.

This shift is part of a broader evolution known as the agentic web. Instead of browsing for answers, consumers are increasingly relying on AI to filter, decide and even act for them when it comes to search. Agentic means being capable of achieving outcomes independently, or simply put, acting as an agent.

To visualise the difference:

• Before: You‘re looking for new trainers on a budget, so search for “best trainers under £100” and browse for 20 minutes through 10 websites that were SEO-optimised to appear at the top of the results, before choosing one that catches your eye
• Now: You ask an AI agent to “find you the best trainers under £100” and get the answer you need after the AI agent checks reviews, compares prices, and finds you a pair of trainers you can order (or in some cases, the agent can order) with no browsing required

It’s seamless and efficient for consumers, but for publishers and advertisers, it collapses the traditional discovery funnel.

An existential threat

Publishers are already feeling the pain. The Professional Publishers’ Association (PPA), which advocates for UK publishers, has described the rise of AI search features as an “existential” challenge, warning that AI Overviews and Google’s new AI Mode are undermining referral traffic by turning search visibility into citation with no engagement.

Meanwhile, Google is disputing the claim, insisting traffic remains stable and now focuses on quality, but without providing any hard data to prove it. (Read more in: The Media Leader)

The backlash isn’t confined to industry bodies. The UK’s Competition and Markets Authority (CMA) has proposed to designate Google with ‘Strategic Market Status’ (SMS) in general search and search advertising, a move that enables imposing conduct requirements on Google to address competition concerns, including those related to AI search features.

But in the meantime, publishers and advertisers have to take a new tack. This is where GEO comes in.

GEO: a lifeline?

GEO, or Generative Engine Optimisation, means to optimise your website for AI engines and platforms. The GEO conversation is still in early days and how to best go about it remains to be seen, but here’s what practical GEO might look like:

• Improving clarity, structure, and authority,  all of which   are rewarded by AI systems. Content that is vague, clickbaity or poorly sourced is less likely to be cited by AI. LLMs like Google’s AI Overviews or ChatGPT are trained to give privilege to content that comes from identifiable, credible humans and web domains over anonymous or generic pages.
• Exploring partnerships and licensing could be a solution. Some publishers are licensing content to AI platforms (such as Associated Press and OpenAI) while others are experimenting with APIs to provide structured data directly to AI systems, at cost.
• Rethinking ad strategy. If fewer users are clicking through, advertisers need to meet them earlier in the funnel, inside conversational interfaces, product recommendation agents, or branded AI tools.
• Keep an eye on MCP (Model Context Protocol). This was developed and released in open-source in November 2024. It standardises how an AI application connects to external sources, acting as middleware for data reaching LLMs. In theory, it could help users access publishers’ content in the AI tools of their choosing. More on this on Digiday.

The big concerns

Optimising your content for AI sounds a little bleak in an industry that’s fighting to stay connected and human-first. And there are other big unanswered questions too. GEO runs a high risk of being inequitable, giving the biggest brands and publishers all the power whilst smaller players have less and less chance of being visible.

On top of that, the funding model for quality journalism is at risk. AI engines are scraping freely available content without compensation, and quite often without attribution. How can publishers monetise their ad space to support their platforms when that ad space is rapidly losing value? Models like MCP offer some solutions, but the bar is getting higher.

And it raises concerns not just financially, but on regulation as well. Intellectual property and usage rights have to be called into question. Will regulators push for more than transparency? What about content licensing, revenue sharing? 

Whether AI is copyright infringement is another whole debate, but the danger here is that the web is being hollowed out. If AI systems feed off content that is no longer economically sustainable to produce, then without intervention, there’ll eventually be no more content to feed off. The ecosystem risks cannibalising itself with contradictions.

Adapt to survive

One thing is for certain: nothing in this scenario is future gazing. Search is changing fast, shaking up a long-standing system of visibility and value. AI and the agentic web are rewriting the game. While regulators like the CMA are taking steps, publishers and advertisers shouldn’t wait for rescue. Although SEO won’t disappear overnight, the data shows that its value is diminishing by the day.

But adapting to survive is something the advertising industry has always done. Something humans have always done. This is far from the first time the industry has faced disruption like this.

From the dawn of programmatic to constantly evolving channels, adtech has always found a way to evolve and thrive. Zero-click and the agentic web simply represent the next frontier. There’s time now to experiment and collaborate. Push for transparency and play a part in controlling the outcome.

Either way, the advertising challenge remains the same as it always has: how do we ensure content, brands and products get discovered, whatever the environment people are searching in?

Additional Reading

"PPA asks CMA to require greater transparency of Google’s AI search features" – The Media Leader

The Professional Publishers Association requests conduct requirements for Google and other nascent AI search platforms.

"WTF is Model Context Protocol (MCP) and why should publishers care?" – Digiday

Model Context Protocol, or MCP, standardises how an AI application connects to external sources.

"CMA takes first steps to improve competition in search services in the UK" – GOV.UK

The Competition and Markets Authority (CMA) is proposing to designate Google with 'strategic market status' (SMS) in general search and search advertising.

"UK and US publishers say Google AI is harming website traffic" – Press Gazette

Leading US online publishers reported an average fall in referral traffic from Google search of 10% in May and June compared to a year earlier.