The Shelf Moved Inside the Answer

Marketing organizations are splitting into two camps on artificial intelligence, and the division is not between adopters and holdouts. It runs between teams automating internal workflows and teams fighting for visibility inside AI answer engines, a discipline now traveling under two competing names, generative engine optimization and answer engine optimization.

Adweek’s latest round of interviews with enterprise marketers makes the split concrete. Bayer is building discoverability and shopability across six over-the-counter brands inside AI answer engines. Zoom automated product naming with an internal agent and is hiring a dedicated AI engineer to extend the approach. A health tech senior vice president is running Profound and Writer in combination to scale discoverability and content output simultaneously.

Those are three different problems wearing the same label, which is the most useful thing about the sample. Agentic transformation is not one initiative. It is a set of unrelated bets being placed at different speeds by companies with different exposure.

Bayer Goes After Six Brands at Once

The Bayer approach is the most instructive because the category makes the stakes visible. Over-the-counter health products are bought through a question. Someone asks what to take for a symptom, and for twenty years that question went to a search engine that returned ranked links, several of which a brand could occupy through paid placement.

That question now increasingly goes to an assistant that returns one synthesized answer citing a handful of sources. Similarweb’s 2026 brand visibility index puts 35% of U.S. consumers using AI tools at the product discovery stage against 13.6% using traditional search. Ahrefs data shows the average site’s search traffic down roughly 21% while AI-referred traffic multiplied roughly tenfold. Boston Consulting Group has tracked the zero-click pattern for years, with close to 60% of searches ending without a click.

Pursuing shopability rather than mere mention is the sharper part of Bayer’s strategy. Appearing in an answer is a brand awareness outcome. Being the product an assistant can route a user to purchase is a distribution outcome, and the difference between them will define which companies treat this as a communications problem and which treat it as a channel problem.

Zoom Automates the Work Nobody Wants

Zoom’s product naming agent is a smaller story with a broader application. Naming is slow, consensus-driven, and consumes senior attention disproportionate to its strategic weight. Handing it to an internal agent removes a recurring bottleneck without touching the work marketers actually want to protect.

The hiring signal matters more than the use case. A marketing organization recruiting a dedicated AI engineer has stopped treating this as a software procurement question and started treating it as an engineering capability it needs in-house. That is the point where a company’s automation stops resembling its competitors’, because the systems become specific to its own data and workflows.

Most teams will never hire an AI engineer. The transferable principle is choosing automation targets by how much senior time they consume rather than by how impressive they look in a board deck.

The Tooling Arrived Before the Standards

The vendor landscape has expanded faster than the practice. By mid-2026, at least eight platforms competed to measure brand visibility inside AI answers, according to landscape analysis published by CiteLens. Profound anchors the enterprise end with panel data on what users actually ask AI systems and analytics tracking how AI crawlers read a site.

Platform-side reporting is arriving too. Microsoft surfaces AI performance data through Bing Webmaster Tools, including citations and cited URLs. Google began testing dedicated generative AI performance reports in Search Console in June.

The early evidence on what moves visibility contradicts the SEO instincts most teams carry. Brand mentions correlate substantially more strongly with AI visibility than backlinks do. Distributing content across a range of publications has been associated with citation increases several times higher than publishing only on an owned domain. Both findings point the same direction. Earned presence across the open web matters more to a language model than technical authority concentrated on one property.

The Measurement Gap

The honest position on this discipline is that measurement remains its weakest component. AI responses vary between sessions, across engines, and over time, which means a movement in visibility can reflect ordinary response variance rather than any change a team made. Testing that survives scrutiny requires repeated prompts, multiple phrasings, and results separated by engine.

There is also a definitional problem nobody has solved. A citation is not a click, a click is not a visit, and a visit is not attributable revenue. Teams reporting share of AI voice to executives should be explicit that they are reporting a proxy, because the gap between proxy and outcome is where credibility is lost when budgets are reviewed.

What Lean Teams Should Do First

For a small marketing organization, the practical entry point is not a platform purchase. It is a baseline. Write out the twenty to thirty questions a prospective customer would ask an assistant before buying in the category, run them across ChatGPT, Gemini, Perplexity and Claude, and record which brands and sources appear. That exercise costs an afternoon and produces a competitive map most companies have never seen.

The second step is auditing whether the content that should answer those questions exists in a form a model can retrieve. Original data, clear structure, explicit recency signals and specific claims are cited more readily than general commentary, which is the same content quality argument marketers have been making for a decade with a new mechanism enforcing it.

The third is distribution. If earned mentions across the open web carry more weight than owned-domain authority, then digital PR and publication partnerships move from a brand-building line item to a discoverability requirement. That reallocation is the uncomfortable part, because it takes budget from a channel most teams can measure and moves it to one they cannot measure yet.