A monitoring vendor shipped a free, no-login AI visibility report yesterday that shows the customer the real answer text across three engines, which is three quarters of our free check given away instantly.

Reported verbatim: "Google is testing a thinner design for the citations/link panel on the right side of the AI Overview." Spotted by Khushal Bherwani on X, who described it as a different style for the reference panel. There is no Google comment and no indication of test scale. Barry Schwartz's own line, verbatim: "I am not sure if making links and citations is a good strategy to make the publishing world love you, Google."

A second interface test landed the same morning. AI Mode is being shown with paginated follow up responses and a skip button, spotted by Sachin Patel, described as: "After submitting an answer, AI Mode generated a new query within the same chat based on the response." Nothing in that report touches citations.

Mueller on programmatic SEO, verbatim: it "often leads to a site that's either spam, borderline spam, or low quality", and "It's easy to spin something up with many pages, it's hard to provide real value to users." On the consequence, verbatim: "Our systems have possibly lost faith in your site providing good value to users based on the old pages." On recovery, verbatim: "Resolving this tends to take time & significant effort to show the value."

Older material, never carried here, and it is the parent of the 3 August study already on this desk (the 74% cited versus 44% named finding). Same tracking, different question. Method stated on the page: 1,094 US categories tracked monthly in ChatGPT from January through June 2026, five representative buyer prompts per category, over 50,000 brands, over 220,000 domains, over 220,000 URLs and over 600,000 citations. An owner is defined verbatim as "the brand with the highest share of mentions, named in at least four of five prompts, with at least a five-percentage-point lead over the runner-up."

The figures, verbatim: "Only 15.2% of the 1,094 categories we analyzed had a clear owner", and "85% remain open to competitors that can build consistent visibility". Split by category size: "Only 11.3% of the top-half topics had a clear brand owner", against 19% in the bottom half. On stability: "Owners with a 5-percentage-point lead retained first place in 90% of month-over-month comparisons", with leadership changing hands regularly in categories where the lead was small. And on the link to conventional metrics: "Owners had higher branded search volume in 55.7% of pairs", which is barely better than a coin toss. Indig's own caution, verbatim: "Treat that as a hypothesis, not a finding. What the data actually shows is: traditional SEO metrics aren't enough to explain who owns a topic."

Limits that must travel with it. It is one engine, ChatGPT, of the five we run. The page does not state runs per prompt or whether capture was through the API or the consumer interface, which are the two facts that decide reproducibility. The categories are commercial product and service categories, not coaching. Semrush sells an AI visibility product. And the authors say plainly that the data cannot explain why leadership switches: "Why the switches happen isn't something this data can answer."