AI search shifts from retrieval to delegation, reshaping brand choice

From retrieval to delegation

Becky Simms’ delegation-search framing gets at the real change: people are no longer using search only to retrieve information, they are asking AI to do the comparison work for them. That shift shows up everywhere from travel planning to product research, where the goal is less “show me links” and more “tell me what to choose, and why.”

That matters because delegation changes the unit of value. In the old model, a page could win traffic by being easy to find. In the new one, a brand has to be easy for an assistant to understand, summarize, trust, and recommend. If an AI system can confidently give the user a short list, a best option, or a next step, the brand that gets chosen may be the one with the clearest proof, not the loudest SEO footprint.

Why the click is shrinking

The pressure on agencies is not theoretical. Pew Research Center reported that in March 2025, 58% of respondents did at least one Google search that produced an AI-generated summary. In that same analysis, users were less likely to click links when a summary appeared, and they very rarely clicked the cited sources. That is the practical reality of the new search interface: the answer is increasingly consumed before the visit ever happens.

The long-standing zero-click pattern backs that up. SparkToro’s 2024 study found that in the United States, only 360 of every 1,000 Google searches sent a click to the open web. In the European Union, the figure was 374 out of 1,000. Similarweb has also pointed to the same direction of travel, with AI Overviews and other SERP features taking a larger share of attention that used to flow to blue links.

For agencies, that means the KPI conversation has to move earlier in the funnel. If a client is still measuring success only by session starts, rankings, and last-click conversions, they are already measuring too late for the way AI-assisted search is actually behaving.

What brands have to be good at now

Simms’ main point is not that content disappears. It is that content has to do a different job. It must help AI systems recommend the brand at the moment of decision, which means the page needs cleaner positioning, more explicit proof, and a structure that can be summarized without confusion.

That is especially important for brands in competitive categories where users are comparing options. In a delegated search flow, the winning asset is often not the most expansive one, but the most legible one: a page that makes pricing, features, tradeoffs, service area, guarantees, reviews, and use cases easy to parse. If the content is vague, padded, or overloaded with marketing language, it becomes harder for an assistant to turn it into a confident recommendation.

Agencies need to think in terms of brand signals as well as content. Reviews, case studies, expert bios, product specs, service details, and third-party validation all become part of the decision surface. The job is to make the brand look like the safest answer when an AI tool is deciding which option to put in front of the user.

The shape of content has to change

The practical rewrite starts with clarity. Pages that used to be built mainly for keyword coverage now need to answer the comparison questions that humans and AI assistants both ask: What does this do? Who is it for? What makes it better? What is the evidence? What is the downside?

That is where structured proof matters. Agencies should be packaging information so it can be extracted cleanly, whether that is through schema, tightly written service pages, comparison tables, review snippets, or plainly stated differentiators. The point is not to stuff more keywords onto the page; it is to make the brand easier to cite, summarize, and rank in an AI-mediated choice.

The best content systems will also be modular. A single article, landing page, or product page should be able to support multiple journeys: early-stage education, side-by-side comparison, and final selection. That is how you stay useful when the user is no longer just searching, but delegating.

Google is already telling site owners where this goes

Google’s own documentation now reflects the shift. Its Search guidance explains how AI features such as AI Overviews and AI Mode work from a site owner’s perspective, and it includes direction on how to approach inclusion and measure performance in those experiences. That is a meaningful signal: the platform itself is not treating AI as a side experiment, but as part of the core search product.

Google said in August 2025 that AI in Search was driving “the most significant upgrade” to Search and also emphasized that it still sends billions of clicks to the web every day. Those two messages can coexist. Search is evolving, but the web still matters, which means the brands that win will be the ones that help AI features surface them accurately without relying on traffic alone.

For agencies, that should sharpen the pitch. The question is no longer just whether a page can rank. The question is whether the brand is legible enough for Google’s AI systems, and persuasive enough for the user who may never read beyond the summary.

Measurement has to move beyond the pageview

This is where many teams will feel the pain first. If AI is acting as a decision layer, the old attribution chain gets thinner, and some of the influence happens before a visit is recorded. That does not make measurement impossible, but it does mean agencies have to start treating visibility, mention quality, and inclusion in AI-shaped experiences as real performance signals.

Google’s own documentation points site owners toward measuring performance in AI features, which is exactly the kind of shift teams need. Search Console and similar reporting layers become more important when the question is not simply “Did the user click?” but “Did the brand show up in the right way inside the AI-assisted journey?”

That is also why earlier-funnel influence matters more than ever. If a user gets confidence from an AI-generated summary, the brand may have won the decision before the site visit. Agencies that can explain that to clients will have a much stronger story than those still trying to defend every drop in organic sessions as if nothing in the interface had changed.

The agency opportunity is now a service-model shift

McKinsey’s 2025 State of AI survey found that organizations are increasingly using AI agents, but most are still early in scaling AI for enterprise value. That is exactly where agencies can matter most. Clients need help translating the rise of agentic behavior into content systems, brand proof, and measurement models that actually work in delegated journeys.

Semrush’s 2025 findings point to the same commercial pressure. As commercial and transactional queries increasingly trigger AI Overviews, and navigational queries do too, the old boundary between discovery and decision keeps collapsing. Agencies that understand that overlap can sell more than SEO execution. They can sell recommendation readiness.

That is the real takeaway from this shift: the win is no longer just being found. It is being chosen on behalf of the user, by a system that has already decided what counts as credible. The agencies that adapt fastest will be the ones that build for that moment with sharper content, stronger brand signals, and proof that an AI can trust.