Stay up to date with the latest insights on AI visibility, GEO optimization, and digital transformation.
Showing 1-9 of 23 posts
Your AI search numbers are only as good as the questions behind them. A step-by-step method for building a 30 to 50 question set from sales calls, search data and reviews, balancing it across buying stages, and keeping it stable enough to show trends.
On September 16, 2026, OpenAI began testing Sponsored Agents, an ad format that lets a ChatGPT user click an ad and open a labeled conversation with an agent the advertiser sponsors. Here is what OpenAI actually announced, how it changes the ways a brand can appear inside ChatGPT, what the answer-independence promise does and does not cover, and what B2B marketers should do now.
On October 8, 2026, Anthropic published a Usage Policy that bars using Claude to seed the sources search engines and AI systems draw answers from with content that misrepresents its origin, authorship or independence. Two weeks earlier Microsoft rebuilt Copilot around work. What both changes mean for how B2B brands earn visibility in Claude and Copilot.
Brands tend to appear when buyers name them and vanish when buyers ask category questions. Here is how an AI answer to a vendor question comes together, the five places B2B brands drop out, and what to fix first.
A buyer asks an AI assistant to compare you with three alternatives and hears about the company you were a year ago, not the one you built. Some of that is model error. Some of it is a real gap between what you claim, what you deliver, and what the market can verify. How to tell the difference, and what to fix at the source.
AI visibility gets muddled when citation, mention and recommendation are treated as the same thing. They are different outcomes, they diverge sharply by engine and query type, and recommendation sits closest to the buyer's consideration set. How to define and count each one, and how breaking them down by engine, topic, buyer stage and persona turns AEO into market research.
The Common Crawl Foundation analyzed every llms.txt file in its July 2026 crawl archive. 68.27% were plugin or template output, Wix alone accounted for 41.34%, and 22.56% contained no links at all. Here is what that means, what it does not mean, and what the evidence actually supports doing for machine readability.
A 41-day server-log experiment found GPTBot, ClaudeBot, PerplexityBot, Meta-ExternalAgent and Amazonbot discovered none of 293 JavaScript-linked pages. Google found 48%. After conversion to plain HTML, GPTBot crawled 250 of them in under three hours. What that means for crawl access, robots.txt and the metrics you should be watching.
AI engines describe 96% of brands accurately and never mention 89% of them in category answers. One industry study puts 82% of AI citations on earned media. Brands with fewer than 2,000 pages mentioning them surfaced 3% of the time. The constraint on AI visibility is not your website.