Google Search Is Becoming Personal and Generative

Visibility Report #31 | August 25, 2026

Google is giving users more control over the sources they see while AI Mode is pushing queries and interfaces farther from the old blue-link model. The operating question is no longer just whether a page ranks. It is whether a person recognizes the source, whether an AI system can extract the answer, and whether the interface leaves any reason to click.

Google Expands Source Personalization

Google announced new controls for personalizing publishers and topics across Search, Discover, and News. The immediate impact is audience loyalty: users who actively choose a source may see more from that source in future discovery surfaces. The AI-search impact is less settled. Google has not said this new control changes AI Overview or AI Mode citation behavior, so this should be tracked as a source-recognition signal, not treated as an AI ranking lever.

Start with the basics: make the publisher, author, topic focus, and editorial promise unmistakable. A user cannot prefer a source they do not remember, and an answer engine has less to work with when brand and expertise signals are muddy.

📎 Google Search Blog​

AI Mode Queries Are Reportedly Three Times Longer

Search Engine Journal reports that AI Mode queries are roughly three times longer than conventional searches. That means more searches may arrive as full buyer questions, with context, constraints, and follow-ups baked in. The content lesson is not to write longer introductions. It is to answer the specific question earlier and support that answer with evidence that can be extracted cleanly.

Test priority pages against the questions real prospects ask when they already understand the category. If the page only performs for short keywords, it may be underbuilt for the prompts that matter inside AI Mode.

📎 Search Engine Journal​

Generated Interfaces Could Replace Tool Pages

Google is testing generated interfaces inside AI Overviews, according to Search Engine Journal. That matters for any page whose value is an interaction: calculators, comparison tools, configurators, estimators, and other utility pages. If Google can recreate the interaction in the answer surface, the page may still influence the answer without earning the visit.

The smart move is baseline measurement before this spreads. Record impressions, clicks, conversions, query classes, and AI Overview exposure for high-value tool pages now, then watch which interactions Google starts recreating.

📎 Search Engine Journal​

ChatGPT Ads Expand Across Europe

OpenAI says ChatGPT Ads will expand to 31 European countries, extending the paid visibility layer around conversational search. This is not the same story as the earlier product-carousel and bidding updates. It is a market-expansion signal: more brands will need to separate paid ChatGPT exposure from earned recommendations and citations.

For affected markets, capture prompt-level screenshots, referral data, ad presence, recommendations, and cited sources before campaigns begin. Otherwise the first readout will blur paid exposure, organic recommendation changes, and normal answer volatility.

📎 OpenAI​

Information Gain Needs Real Audience Evidence

Ahrefs argues that information gain is useful even if Google does not use a literal score by that name. SparkToro's new persona tool points at the same issue from another angle: thin consensus content gives search engines and answer engines little reason to cite you. Distinctive audience evidence, original examples, and a clear method create source value that rewrites do not.

Pick one commercial page and look for claims only your organization can support. Add the data, method, limitation, and practical implication beside the claim. That is more useful than another paragraph saying the same thing competitors already say.

📎 Ahrefs | SparkToro​

YouTube Keeps Showing Up in AI Visibility Research

Foundation reports that YouTube appeared in 80.2% of the demo searches it studied, while many B2B companies keep product demos only on private video platforms or their own sites. The useful takeaway is not "move every video to YouTube." It is that public distribution, searchable transcripts, and platform-level discoverability can change whether an AI system can find the demonstration.

Treat the 80.2% figure as a directional benchmark, not a universal rate. Test a small set of useful demos on YouTube, then compare citations, branded prompt coverage, referral traffic, and assisted conversions.

📎 Foundation Inc​

More From This Week

Search & AI Visibility

• ​Reddit fades from ChatGPT citations (Axios) - Promptwatch data shows how quickly AI citation sources can shift, which argues against building GEO plans around one borrowed authority surface.
• ​LLM visibility starts with better internal communication (Search Engine Land) - Product facts, customer language, and measurement data need to move across teams before content can answer AI-era questions well.
• ​SEO and Ads Only Work If Your Website Does (Search Influence) - Acquisition work only compounds when the landing experience supports the visit.
• ​The Reddit Cloud 100 (Foundation Inc) - A directional view of where community discussions become answer-engine sources for B2B SaaS brands.

From the Tool Blogs

• seoClarity: ArcAI Indexer - A product update focused on getting new and changed pages discovered faster by AI search systems.
• seoClarity: What People Ask About Your Brand In AI Search - Prompt discovery framed around the questions users ask about a brand inside AI search.
• OtterlyAI: Agent Analytics - Adds visibility into AI agents and crawlers visiting a site, then connects that activity to AI-answer visibility.
• OtterlyAI: Brand Sentiment Tracking in AI Search - Measures whether AI search engines describe a brand positively or negatively across ChatGPT, Perplexity, Gemini, and AI Overviews.
• AirOps: The Top 7 AI Search Metrics for 2026 - A measurement framework for moving beyond clicks and rankings into citations, mentions, and AI-response visibility.
• BrightEdge: The Customers You Are Turning Away Without Knowing It - BrightEdge says AI-agent traffic is approaching human organic search volume for some sites, which makes bot experience a real measurement concern.
• Profound: YouTube Citation Breakdown - Adds structured reporting for videos and channels that appear as AI-answer sources.
• Profound: Where Do Inaccurate AI Claims Come From? - Uses FactCheck data to examine where incorrect brand claims originate and spread.
• Profound: Commercial Conversations in ChatGPT More Than Doubled in a Year - Vendor research on how often ChatGPT conversations move into product and purchase territory.

Agency & Practitioner Insights

• ​Duane Forrester: AI is rewriting SEO, but the mission is the same (Search Engine Land) - A useful practitioner reminder that AI-search work still has to serve users, build trust, and improve the underlying experience.

Engine Announcements

• ​Five new learning features in Search (Google) - Study tools are another example of Search turning into a task surface, not just a result list.

So What Do You Do About It?

Choose five commercial questions that matter to your business. Record the answer, citations, ads, source types, and any generated interface shown by Google AI Mode and ChatGPT, then repeat the same test weekly for a month. Analytics are not truth. They are opinions with decimal points, but a consistent prompt set can still show whether your source visibility is improving or just moving between surfaces.

The Visibility Report | Will Scott​
This newsletter is produced collaboratively by Will Scott and Bob, an AI agent. Human oversight, AI efficiency.
​Unsubscribe​