A Working Framework for AI Search Shifts You Can’t Afford to Ignore

AI search is changing at a pace none of us has experienced before in marketing. The presentations I saw at Zero Click NY highlighted both how much AI search has changed over the past six months and the characteristics that may become lasting features of the landscape.

Of all the points covered, these seven stood out as the most important. From the rise of the marketing engineer, to the differences between Claude and ChatGPT results, to Claude's meteoric rise among businesses over the past 12 months, here are the most impactful takeaways I left with.

1. Every AI relies on different content

ChatGPT and Claude share only about 8% of their citations, per Profound data. Put differently, 92% of what ChatGPT cites wouldn't be cited by Claude for the same query. A brand can own visibility in one engine and be virtually invisible in. For search teams, the important part is not the headline movement by itself. It is whether the shift changes which communities, forums, video surfaces, or publisher pages now satisfy the query better than the old ranking pattern. The same pattern also shows up in 4 Layer AI Ops Playbook, where the practical question is how the signal becomes visible.

The useful check is whether this improves the system behind search performance, not only the words on the page. Internal links, crawlable content, clear entities, current evidence, and a sensible page structure all help the recommendation become easier to trust.

2. Claude is quietly winning B2B, so sequence your optimization

If you've seen the generative AI traffic share charts, Claude looks like a rounding error. But web traffic is the wrong chart. Roughly 85% of Anthropic's revenue comes from enterprise and API usage that never shows up in consumer traffic. The practical read is that brand signals need to be consistent enough for both people and AI systems to form a stable view of the company, its expertise, and its trust signals.

The practical value is in connecting the idea to an observable signal. That means deciding what should be checked, what would prove the issue is real, and where the team should make the smallest useful improvement first.

3. ChatGPT ads are here, and this is what we're seeing

The moment is here: Your competitors are buying visibility through ChatGPT ads. ChatGPT ads are live and self serve, sitting directly inside the chat product. The same two weeks brought GPT 5.5, citation chips turning into clickable. The search implication is whether the section improves the evidence around the page, not simply whether it adds more wording. Clear entities, crawlable structure, internal links, and useful context are what make the topic easier to evaluate.

ChatGPT Ads match on topic

Ads match on topic similarity, not intent. Only 14% of real user prompts carry commercial intent, but 20% of prompts trigger ads, a math problem can serve an ad. The embedding analysis found that ad titles and descriptions are the single. The practical question is what this changes in the system: the page structure, the evidence presented, the measurement habit, or the way the topic is connected to related work.

Paying for ads

"Pay to play" is here. About one in five ad placements appears against a mention of a direct competitor, and the brand mentioned organically shows up as the advertiser only about 8% of the time. Someone else is twice as likely to be the. The practical read is that brand signals need to be consistent enough for both people and AI systems to form a stable view of the company, its expertise, and its trust signals.

Ad inventory is scarce and expensive

ChatGPT shows roughly one ad per conversation, the median conversation is three turns, only 30% of eligible users see ads at all, and CPMs/CPCs are running around four times Meta's. Expect that to change in predictable ways: more ad slots. The practical read is that brand signals need to be consistent enough for both people and AI systems to form a stable view of the company, its expertise, and its trust signals.

4. Claude is the most directly optimizable AI right now

When Claude searches the web, it pulls from Brave. Not "influenced by" Brave. According to the talk I saw, it pulls directly from it. In Profound's latest testing, 79.2% of Claude's citations came directly from Brave's top 10 results for. The search implication is whether the section improves the evidence around the page, not simply whether it adds more wording. Clear entities, crawlable structure, internal links, and useful context are what make the topic easier to evaluate.

5. Claude only performs web searches a third of the time

There's a catch, and it's a big one. ChatGPT triggers web search on roughly 95% of prompts. Claude searches only about a third of the time, likely because every search costs money (Brave's public API pricing runs around $5 per thousand. Local visibility depends on whether the details across pages, profiles, categories, reviews, photos, and service descriptions reinforce the same answer for a specific location based query. This connects with AI Is Merging Paid and Organic Visibility when the same signal needs a clearer operating decision.

The risk is usually hidden in the execution layer. A page can look fine to a human and still fail for an automated visitor if the form, call to action, rendering path, or confirmation step is not accessible enough for the agent to complete the task.

6. Query fan out: A raffle on one stage, near deterministic on another

Two speakers described the same mechanism in almost opposite terms, and the tension between them is instructive. Query fan out is the set of synthetic queries an AI engine runs in the background to gather content before generating an. The search implication is whether the section improves the evidence around the page, not simply whether it adds more wording. Clear entities, crawlable structure, internal links, and useful context are what make the topic easier to evaluate. A useful companion note is AI Recommendation Sets Leave Some Brands Out, because it looks at a nearby part of the same system.

7. The marketing engineer is here, and agents are the new workforce

It would be easy to dismiss "marketing engineer" as a vendor manufactured job title. The hiring market says otherwise. Google has hired its first marketing engineer. Figma posted the role at a $295,000 base salary. RBC and Autodesk have. The measurement question is whether this signal changes a decision, not whether it adds another number to a dashboard. Useful reporting connects visibility, engagement, and business outcomes without pretending every AI influenced journey will produce a clean click path.

The reporting question is whether this signal changes a decision. If it only creates another number in a dashboard, it adds noise. If it helps separate profile activity, website visits, calls, bookings, and direction requests, it can make local performance easier to understand.

What the visibility signal actually changes

What the visibility signal actually changes: a Working Framework for AI Search Shifts You Can’t Afford to Ignore should be treated as a visibility signal, not a standalone headline. Introduction AI search is changing at a pace none of us has experienced before in marketing. The presentations I saw at Zero Click NY highlighted both how much AI search has changed over the past six months and the characteristics that may become lasting.

What the visibility signal actually changes: the practical question is whether the page, brand evidence, and surrounding content make the answer easier to trust. If that support is weak, search systems can still understand the topic but fail to connect it confidently to the brand.

What the visibility signal actually changes: that is why the response should begin with an audit of the evidence already on the site before creating a new asset. The fastest improvement is often a clearer page, a better internal link, or a stronger explanation of why the brand belongs in the answer.