Google's CEO has made something explicit that many in search have been circling around: informational queries are becoming agentic, and search itself will function as an agent manager. That is not a distant forecast. It is a description of a direction already underway, with clear implications for how brands get found, selected, and recommended by systems that increasingly act on behalf of users rather than simply responding to them.
Most brands are still thinking about search in terms of rankings and click-through rates. That mental model is becoming inadequate. The shift to agentic search means the fundamental transaction is changing - from a user typing a query and choosing a result, to an agent interpreting intent, sourcing information, and taking action without the user ever seeing a list of options.
What 'Agent Manager' Actually Means in Practice
When Google's CEO describes search as an agent manager, the implication is that Google will coordinate multiple AI agents to fulfil a user's goal. A user asking how to book a business trip, find a supplier, or compare financial products would not be presented with ten blue links. An orchestrating layer - search as agent manager - would delegate tasks to specialist agents, synthesise outputs, and surface a result or complete an action directly.
This is not speculative design fiction. Google AI Overviews already remove the click for many informational queries. Perplexity and ChatGPT handle research tasks end-to-end. The agentic model extends this further: instead of a single AI response, you get a coordinated pipeline of AI actions. The user's interaction point moves further up the chain, and brand exposure moves further downstream - buried inside a process rather than visible at the surface.
For marketers, this creates a genuinely new problem. You cannot optimise for a click that does not exist. You cannot bid on a query that is handled internally by an agent pipeline. What you can do is influence whether your brand's information, products, and structured data are the inputs those agents prefer to use.
The Shift from Query Matching to Agent Trust
Traditional SEO is built around matching content to queries. Agentic search requires something different: becoming a source that AI agents trust, return to, and cite when completing tasks. That is a higher bar. An agent acting on a user's behalf is not browsing - it is selecting. The criteria for selection are authority, clarity, structured accessibility, and consistency across sources.
Brands that have invested in entity-based content - clear factual claims, well-structured product and service information, consistent brand data across the web - are better positioned for this model. Brands that have relied on high-volume thin content or paid link acquisition are not. The agent does not reward SEO tricks; it rewards information quality.
This is why Generative Engine Optimisation matters now, before agentic search is fully mainstream. The patterns being established in how AI Overviews and LLM-based tools cite brands today are likely to carry forward into more agentic systems. Getting cited in a ChatGPT response in 2026 is practice for being selected by a Google agent pipeline in 2027.
Paid Search in an Agentic World
Google's own CEO framing search as agentic has direct implications for how paid search works. If the user-facing search interface is increasingly replaced by agent-mediated interactions, the inventory available for traditional keyword-based advertising shrinks. This is part of why Google has been moving so aggressively toward Performance Max, Demand Gen, and AI MAX - campaign types that are not dependent on a specific query being typed into a search box.
Performance Max already operates across inventory that extends well beyond the search results page. Demand Gen targets users in discovery contexts - YouTube, Gmail, Discover - where intent is inferred rather than explicit. These formats make more sense in a world where the search results page is no longer the primary point of contact between user and brand. Google is, in effect, preparing its advertising products for the same shift its CEO is describing in search.
For UK advertisers, this reinforces the case for moving budget and strategic attention toward audience-signal-based campaigns rather than over-investing in exact match keyword defence. If agentic search reduces the volume of exposed queries, keyword-centric campaign structures become more fragile. Broader, AI-managed campaign types become the practical alternative - not because they are fashionable, but because the inventory they access is more durable.
What Brands Need to Do Before Agents Are Everywhere
The window to establish AI-visible brand presence is still open, but it is narrowing. When agentic search is the default rather than the exception, the brands that agents have learned to trust will have a structural advantage that is difficult to displace. This is similar to how brands that earned strong domain authority in the early 2010s retained ranking advantages for years - except the mechanism is different and the criteria are shifting.
Practically, this means auditing how your brand appears in AI-generated responses right now. Test your brand across Google AI Overviews, Perplexity, and ChatGPT for the queries most relevant to your category. Note where you appear, where competitors appear instead, and what sources AI tools cite when they do mention you. That data tells you what agents are already learning about your brand.
Then work backwards. If AI tools are not citing your brand for product category queries, that is a content and authority gap. If they cite you but get key facts wrong, that is a structured data and schema problem. If competitors are consistently preferred, look at what their content does differently - how claims are structured, what third-party sources corroborate them, how comprehensively they cover the topic. These are solvable problems, but they require treating AI visibility as a distinct discipline rather than an extension of traditional SEO.
The Strategic Implication Is Already Here
Google's CEO describing search as a future agent manager is notable not because it is surprising, but because it makes the direction official. The informational query - the backbone of organic search traffic for two decades - is being systematically replaced by AI-mediated responses. That process is already underway. The full agentic model, where search coordinates multiple agents on a user's behalf, is the next stage of something already in motion.
For brands, the question is not whether to respond to this shift but how quickly. Waiting for agentic search to be fully mainstream before investing in AI visibility strategy is the same mistake many brands made waiting for mobile, or for voice, or for AI Overviews. By the time the shift is obvious to everyone, the brands that acted earlier will already have the advantage. The CEO of the company building this infrastructure has now told you it is coming. That is a reasonably reliable signal.