Who Owns AI Search? The Wrong Team, at Most Companies

Tom Critchlow said something in a recent conversation with Siege Media’s Ross Hudgens that should make every organic search team a little uncomfortable.

“The people that drive SEO outcomes are not SEO professionals.”

In most organizations, the factors that determine search visibility, including content quality, brand strength, reputation, and mentions across the web, belong to brand, product, PR, and editorial teams. SEO teams may report on the channel, but whether executives believe they can materially influence it is, in Critchlow’s words, “not a given.”

That was already true in classical SEO. It becomes impossible to ignore in AI search, where models assemble much of their understanding of a company from sources the company does not control.

So who should own AI search?

One person should be accountable for it, but no single function can own all the work.

The owner’s job is not to absorb SEO, PR, brand, product, lifecycle, content, and engineering. The job is to make those functions work against a shared understanding of how AI systems represent the company and influence buyer decisions.

At most companies, nobody is accountable yet. The in-house search leaders I interview are trying to assemble cross-functional groups, but many are hitting the same wall: leadership has not named an owner, established a mandate, or committed a budget.The companies furthest ahead usually have an internal champion with enough organizational credibility to convene the necessary teams. Their advantage is not that they have settled the acronym debate or what role SEO plays in GEO, it is that someone has the authority to get the work moving.

Digiday reported in April that Pfizer built an in-house SEO and AI-discoverability team in roughly 60 days. Georgia-Pacific and U.S. Bank had also established internal teams, while Adobe, Hertz, T-Mobile, Lowe’s, and Skims were recruiting leaders with responsibility for SEO and AI discoverability.

Across the most advanced teams I’ve interviewed, a formula for success is being defined. Develop a central team of roughly five to ten people comprised of specialists across marketing, PR, brand, content, and technology. The same pressure is reaching agencies. Digitas has reorganized paid and organic search teams around shared client objectives. The logic is straightforward. When AI changes discovery across every channel at once, separate teams with separate scorecards become harder to defend.

Why no single team can own this

Earned media is one of the clearest levers.

A University of Toronto-led study of four generative engines found that AI platforms systematically favoured earned-media sources over brand-owned and social content.

That puts a significant part of AI visibility within the purview of PR, communications, and brand teams. These teams shape the third-party coverage, expert commentary, reviews, and independent references that AI platforms use to understand whether a company is credible.

Personalization adds another lever.

In a controlled test of Google AI Mode with Personal Intelligence enabled, iPullRank found that appearances of a seeded brand rose from 23.9% to 66.8% when the brand was present in the tester’s Gmail. Gmail was the strongest personal-context signal tested. The experiment does not establish email as a universal ranking factor. It does demonstrate that a company’s existing relationship with a customer can change which brands appear in an AI-generated recommendation.

That lever sits with lifecycle and CRM teams, which have historically had little reason to think of themselves as part of a search program.

The technical layer is real, and it is becoming more complex. Retrieval, chunking, structured content, information architecture, crawler access, and the increasingly agentic ways systems locate and evaluate sources all matter. Much of that work belongs to SEO and engineering.

But Mike King has described the emerging discipline, relevance engineering, as an intersection of information retrieval, UX, AI, content strategy, and digital PR. Most of those capabilities sit outside of conventional SEO teams. Technical SEO may provide part of the foundation, but it cannot independently create the brand signals, external authority, customer relationships, or product clarity that AI platforms rely on.

Measurement also looks less like rank tracking and more like brand tracking.

Critchlow argues that AI-search measurement must account for awareness, preference, consideration, and the qualitative shape of how a company is described. Brand and insights teams have spent decades measuring those questions. Most search teams have not.

That does not mean rank and citation data are irrelevant. It means they are incomplete.

Sentiment scoring is the entry point to that work, and most tools stop there. Even the legacy platforms are pulling a few visibility percentages into a dashboard and calling it AI analytics. The deeper layer is Brand Fidelity, which I’ve coined to mean whether the model’s stored picture of a company, brand, or offering, is accurate.

Critchlow described one of his client’s failures: a brand cited on nearly every query in its space, with the LLM telling people to avoid buying from it. Catching misrepresentations of that magnitude requires measuring all three layers: visibility, reputation with sentiment, and brand fidelity, at the brand and offering levels. Gander is built around these three layers: visibility, reputation, and Brand Fidelity. Each can be measured at the level of the overall brand and at the level of an individual product, service, or offering. Visibility analysis is drawn from uncapped, unfiltered responses so the calculation reflects the full response. Capped retrieval data (specifically fanout queries) or heavily filtered response sets can materially distort visibility metrics. The methodology tools use matter as much as the resulting percentage.

What it looks like when someone actually builds it

The two most effective programs I encountered in research interviews this quarter shared one structural feature: they combined centralized accountability with distributed execution.

At an enterprise HR-software company, an in-house search lead runs a weekly committee spanning SEO, web, content, PR, and the teams responsible for review platforms. Each function owns a defined pillar of work, but all report against shared AI-visibility measures. The company’s self-reported brand-mention rate, measured through its prompt-tracking platform, increased from 15% to 53% over six months. Reviews, previously treated primarily as sales enablement, were reclassified as a performance function after the team found that weak review coverage was limiting citations and recommendations.

A major consumer-finance brand went further. It created a full-time AI-search role and announced the appointment across its marketing and product organizations so every team knew who was accountable. The role was not created to replace SEO, PR, social, affiliate, or product teams. It was specifically created to coordinate them. The person leading the program told me that the important question was not whether SEO or PR should win ownership. The company needed someone who could mobilize the relevant teams and translate shared goals into coordinated projects. His closest partnerships had already shifted beyond technical teams to affiliate, social, and PR.

Note: neither company waited for the industry to settle the acronym debate. They named an owner and got on with the work. Both understand that while SEO does serve as a foundation for AI search, they are absolutely not the same.

Executive attention already exists

Executive attention is usually scarce. For now, a meaningful portion of it is pointed at AI search.The question is whether companies convert that attention into an operating model or allow it to dissolve into organizational silos, competing when they should be collaborating.

Ramp’s spending data shows that paid adoption of visibility tracking tools increased 2.4 times over the preceding year. That demand is also a recurring theme in Critchlow’s writing and in the peer community he runs for senior search professionals. Leadership teams are asking what AI search means for demand, brand visibility, reputation, and competitive position. Jessica Bowman made the case for integrating SEO across the organization years before generative search became a board-level concern.

AI search makes that operating principle harder to avoid. The opportunity now is to make smarter internal and external investments while executive attention is still available.

Critchlow calls a defensive stance in this moment “a career-defining mistake”.

The version of that mistake I keep seeing is endemic. SEOs want to own AI search, while simultaneously stating that it does not meaningfully differ from traditional search. Those positions cannot both be true.

If AI search is simply SEO, there is little reason to build a new organizational model. If it depends on brand, PR, product, lifecycle, content, engineering, and insights, leaving it inside one function’s job description guarantees that most relevant tactics will never executed well.

The answer is to establish a cross-functional team, give it a named owner, assign each function a clear pillar of work, and put shared measures on the same reporting cadence as traffic and demand.

The companies in the top decile of their industry already have.