SEO teams are used to being held accountable for visibility. If a page cannot be crawled, a template produces duplicate content or an important category loses rankings, there is usually a recognizable chain from technical or content diagnosis to an SEO action plan.
AI recommendations break that ownership model. In a new Search Engine Land interview, enterprise SEO veteran Jessica Bowman argues that a brand’s presence in AI answers is increasingly shaped by signals produced across the entire company, including reviews, reputation and operational activity that an SEO team cannot directly control.
The consequence is organizational as much as technical. SEO can make a website understandable, accessible and well documented, but it cannot single-handedly make a company recommendable if the wider evidence available to AI systems points in another direction.
That makes AI visibility less like a conventional SEO deliverable and more like an enterprise capability.
SEO can make a brand understandable without making it recommendable
Bowman has been developing this distinction in her recent Search Engine Land work. In an August analysis, she argues that AI visibility now has two jobs: execute the work SEO and development teams can control, then mobilize the rest of the organization around the factors they cannot.
A company can have a technically sound website. AI crawlers can access its pages. Product information can be structured clearly. The brand can even appear regularly in informational answers and earn citations.
Then a prospective buyer asks which product or provider should actually be chosen for a particular situation, and the brand disappears from the recommendation.
That is because being retrievable is not the same problem as being suitable.
Recommendation queries force AI to evaluate the business, not only the page
An informational query can often be satisfied by finding a page that contains the requested fact. A recommendation query requires comparison.
If a buyer describes an industry, budget, technical requirement, delivery constraint or operational problem, an AI system has to decide which products or companies appear to fit those conditions. That can require information about specifications, service quality, customer experiences, availability and trade-offs that extends beyond optimized marketing pages.
Bowman’s framework is important because it shifts the unit of optimization. The object being evaluated is no longer only a URL. It can be the company and its products as represented across many sources.
Customer reviews can become part of the AI information environment
Reviews illustrate why the SEO team cannot own the entire outcome. Search specialists can improve product pages and structured data, but they cannot directly determine what customers say after a purchase or support interaction.
If public reviews repeatedly describe the same problem, that information can become part of the evidence available on the web when an AI system researches a brand or product. Conversely, consistent positive experiences can create third-party evidence that marketing copy cannot manufacture credibly on its own.
This should not be simplified into a claim that every AI model uses a fixed “review score” as a ranking factor. Providers do not publish a universal weighting formula for recommendations. The practical point is broader: AI systems can retrieve information from sources the SEO team neither writes nor controls.
Operational problems can become visibility problems
Bowman extends the same logic into business operations. Shipping, fulfillment, inventory, support and other operational functions can create public signals that eventually affect how customers and third parties describe the company.
A marketing team can promise fast delivery, but persistent complaints about delays create conflicting evidence. A product page can advertise excellent support, but recurring customer reports about unresolved issues can tell a different story.
Traditional SEO could often treat those operational realities as conversion or customer-experience issues occurring after the search click. AI recommendation systems can move them earlier in the discovery process by synthesizing information before the user ever visits the brand.
In that environment, operations can indirectly shape discoverability because operations shape reputation.
Product teams increasingly influence what search systems can recommend
Product information is another cross-functional dependency. SEO can improve how specifications are exposed, but it cannot decide what the product actually does.
If buyers consistently ask for a capability the product lacks, better optimization cannot make the product a legitimate fit. If a feature exists but documentation is incomplete or inconsistent, the product team and documentation owners may need to solve the information gap before SEO can expose it effectively.
This creates a useful boundary for AI visibility programs. Some failures are discoverability problems; others are product-market or information-quality problems that search practitioners can diagnose but not repair themselves.
PR and reputation are becoming closer to search infrastructure
AI systems also operate across an information ecosystem containing journalism, specialist publications, company announcements, reviews, forums and other third-party material.
That makes public relations relevant to AI visibility for reasons that go beyond traditional link acquisition. Earned coverage can help establish what independent sources say about a company, its expertise and its products.
The same mechanism can work negatively. Controversies, recurring criticism or inaccurate third-party narratives can become part of the source environment that AI systems encounter.
SEO cannot simply rewrite those external pages. Correcting the information environment may require communications, customer support, legal, product or executive involvement depending on the underlying issue.
The old SEO ticket model starts to break
Enterprise SEO has traditionally operated through implementation requests. The SEO team identifies a problem, documents a recommendation and sends a ticket to developers, writers or another owner.
AI visibility can produce issues that do not fit cleanly into a ticket.
“Improve how AI systems recommend us” may depend on better product documentation, fewer customer complaints, clearer positioning, stronger independent coverage and more consistent data across distribution partners. No single Jira task resolves that combination.
The SEO practitioner’s role therefore starts shifting from optimizer toward orchestrator: identify the visibility gap, determine which organizational signals contribute to it and bring the relevant teams into the response.
Being mentioned, cited and recommended are different outcomes
Bowman’s framework also challenges the tendency to reduce AI visibility to one metric.
A brand can be mentioned in an answer without receiving a citation. Its website can be cited as an informational source without the product being recommended. A competitor can receive fewer citations overall while appearing more often when users ask which solution to buy.
Those outcomes represent different stages of AI-mediated discovery.
Technical SEO and authoritative content can be particularly important for retrieval and citation. Recommendation introduces a broader evaluation of fit and reputation. Treating all three as the same KPI can hide the exact problem the organization needs to solve.
Buyer prompts expose weaknesses that keyword reports may miss
Traditional keyword research usually aggregates demand around short search phrases. AI conversations allow buyers to describe much more specific scenarios.
A user can specify company size, industry, compatibility requirements, budget, geography and constraints in one prompt. That gives the AI system more criteria against which to evaluate candidate products.
For brands, these prompts can expose gaps that are not obvious from keyword volume alone. The company may be highly visible for a category term while repeatedly failing to appear when real buyers describe the situations in which the product is supposed to excel.
The appropriate response may involve content, but it may also reveal missing product capabilities, unclear positioning or insufficient evidence from customers and third parties.
AI visibility teams need influence beyond formal authority
Search Engine Land’s AI Brand Visibility Master Class description, led by Bowman, makes the organizational argument explicit. It identifies product information, customer support, documentation, operations, reviews, PR and merchandising among the functions that can contribute signals influencing what AI tells buyers.
Most SEO leaders do not manage all of those teams. They may have no formal authority over any of them.
That means success depends on explaining AI visibility in terms each function understands. Customer support needs to see how recurring issues become public evidence. Product teams need to understand where missing specifications create recommendation gaps. PR needs to see which external narratives AI systems surface. Executives need to understand why visibility cannot be assigned entirely to marketing.
The capability is cross-functional even when the initial diagnosis begins in search.
This does not mean SEO has become less important
The company-wide argument can easily be misread as saying SEO no longer matters. Bowman’s position is closer to the opposite: SEO remains one of the foundational layers, but its successful execution is no longer sufficient for every AI visibility objective.
AI systems still need accessible information. Technical barriers can prevent content from being retrieved. Weak documentation can make products difficult to understand. Poor site architecture can obscure relationships between entities and offerings.
Those are recognizable SEO and web-platform responsibilities.
The difference is that solving them may only get the brand into the candidate set. Whether the brand becomes the recommendation can depend on evidence generated elsewhere.
Company-wide visibility does not imply a known company-wide ranking formula
There is an important methodological caution around this discussion. Bowman’s framework is strategic analysis, not a disclosed ranking specification from Google, OpenAI, Anthropic, Microsoft or another AI provider.
No public evidence establishes a universal formula assigning fixed weights to reviews, operational performance, PR coverage or product documentation across every model and query.
Different systems use different models, retrieval sources and search partners, and their outputs can change between prompts and over time.
Organizations should therefore avoid turning a useful organizational framework into a checklist of imaginary AI ranking factors. The goal is to improve the quality, consistency and availability of the evidence surrounding the brand, then measure how AI outputs respond.
Measurement has to diagnose where the problem lives
A mature AI visibility program should do more than count mentions. It should compare different classes of prompt and inspect why competitors appear where the brand does not.
If the company is absent from informational questions, crawlability, content or authority may be the first places to investigate. If it is cited but not recommended, the problem may sit closer to buyer fit, reputation, product evidence or third-party perception.
If AI answers repeatedly contain inaccurate facts, the team needs to identify the sources producing those facts and determine which department can correct them.
This diagnostic approach gives cross-functional collaboration a concrete purpose. Other teams are not being asked to “do GEO.” They are being asked to solve specific business signals that happen to affect AI-mediated discovery.
AI visibility is becoming an organizational capability
The most important implication of Bowman’s argument is a change in accountability.
SEO can still own technical optimization, content discoverability and much of the measurement layer. It can identify gaps and coordinate responses. But it cannot own customer sentiment, fulfillment performance, product quality, media narratives and every other signal an AI system might encounter.
Holding an SEO team solely responsible for whether an AI assistant recommends the company therefore risks assigning responsibility without control.
The better model is shared responsibility with clear ownership. SEO makes the brand technically discoverable and helps diagnose visibility. Product ensures the offering and its documentation support the claims. Operations and support improve the customer reality behind public feedback. PR and communications strengthen accurate third-party understanding. Leadership makes those teams cooperate when the problem crosses boundaries.
That is a larger mandate than optimizing pages for search engines. As AI becomes a layer between companies and buyers, visibility increasingly reflects the organization that produced the information — not merely the team that optimized the website.