Welcome to Pulse: this week’s updates cover how you measure AI visibility, what to expect in llms.txt, and where the agent web is next.
Microsoft posted new AI citation tools, new data included in llms.txt, Google and two published agent details, and the UK set new Google Search ranking rules.
Here is what is important to you and your work.
Bing Releases AI Citation Share in Webmaster Tools
Microsoft is introducing four new features to the Bing Webmaster AI Tools Performance dashboard. Excerpt Sharing, Objectives, Topics, and Comparisons are all in beta right now.
Important facts: Citation Share reports the percentage of AI citations your site captures for a specific basic query, while Objectives and Topics collects those queries to ease the data limit on the current dashboard. The comparative allows you to put the past tense in the present tense. All four are starting to be released around the world, they are still previews.
Why This Matters
Citation Share is the first metric in Bing Webmaster Tools to show your AI visibility against your competitors, not just whether you’ve been cited. The catch is that the data is only for Bing, which covers Copilot and Bing Answers, so it doesn’t say anything about Google, where Search Console doesn’t provide citation-style statistics.
What SEO Experts Say
Gianluca Fiorelli, Founder of ILoveSEO.net, wrote on LinkedIn:
“Bing Webmaster! The Google Search Console we’d love to have.”
Read our full story: Bing Releases AI Citation Share in Webmaster Tools
Google And Ahrefs Data Is Small Case In llms.txt
llms.txt took two songs this week. Google’s John Mueller said the file can’t help LLM tell one site from another, and new Ahrefs data shows bots are important to download it.
Important facts: Speaking at Search Off the Record, Mueller argued that llms.txt cannot classify sites for discovery, because the file is self-reported by the site in hopes of being selected, and pointed back to plain HTML and internal links instead. Ahrefs data stays in the same place. Out of all 137,000 domains, 97% of llms.txt files yielded zero requests, and retrieval bots that generated citations, such as ChatGPT and Perplexity, made only 1% of downloads that occurred.
Why This Matters
Both results point in the same direction. A self-reported file can’t make LLM pick you, and citation-generating bots can’t download it, so don’t expect llms.txt to drive your AI search visibility. It still gains a narrow niche with coding agents and training crawlers that learn it, which ends up being cheap to maintain, the same conclusion of the SE Ranking that looks at 300,000 domains accessed in the past months.
What SEO Experts Say
Nat Miletic, Founder at Clio Websites, summarized the takeaway from LinkedIn:
“llms.txt is low cost to publish, good to have. Don’t expect it to deliver AI visibility right now.”
Read our full story: Google’s Mueller Says llms.txt Can’t Help LLMs Classify Sites and 97% of llms.txt Files Have No Applications, Ahrefs Data Shows
Google, Microsoft, and Others Publish New AI Agent Details
The two agency clarifications came within days of each other. Google Cloud published the Open Knowledge Format, and a consortium including Google, Microsoft, GitHub, and Hugging Face followed with Agentic Resource Discovery.
Important facts: The Open Knowledge Format, or OKF, is a markup format for packaging organizational knowledge, things like datasets, metrics, and runbooks, in a way that AI agents can read. Agentic Resource Discovery, or ARD, is a well-defined framework for how agents discover and verify tools, capabilities, and other agents. Both are still early, with OKF at version 0.1 and ARD at 0.9.
Why This Matters
Not a single one is asking for you this week. Both repeat the move made by llms.txt, a structured file in your domain for software to read, and the same unresolved discovery question hangs over it. See what formats are available before committing to any.
What SEO Experts Say
Martin Jeffrey, Founder and Chief Strategy Officer at Harton Works, compared ARD to the early days of search on LinkedIn:
“It’s a sitemap, reborn with skills instead of pages.”
Suganthan Mohanadasan, Co-founder at Snippet Digital, who created two free format tools, has strong expectations:
“This is not a magic mushroom and will not increase your AI exposure overnight.”
Read our full story: Google Cloud Announces Open Information Format and Google, Microsoft Back Draft AI Agent Discovery Spec
UK Orders Google Rankings in Search Results
The UK’s Competition and Markets Authority (CMA) has set new rules for Google Search. Google should measure organic results on objective criteria and provide notice before important changes.
Important facts: The Fair Ranking requirement covers UK environmental results, including AI Overview but not ads, and obligates Google to use objective, non-discriminatory criteria, provide advance notice of significant changes, and provide a channel to raise ranking concerns. The second requirement turns Google’s voluntary UK data portability tool into a legal obligation. Google disputed this premise, saying its standard was already fair and transparent.
Why This Matters
Advance notice requirements and complaints are parts that can affect daily work, instead of a black box master update with a warning and a fallback mechanism. The relevant standard also reaches within the AI Overview, and as a requirement to come out in early June for the CMA it only applies to the UK, which has a real impact on how Google chooses to implement it.
What SEO Experts Say
Laura Iancu, Founder of Searchpedia, put it clearly on LinkedIn:
“No more ‘oopsie, we just released another important update.'”
Chloe Smith, Strategic SEO Lead at Blue Array, expects the backlash:
“I expect Google will try to find a way around this.”
Read our full story: Google Must Provide Notice Before Important Ranking Changes
Theme of the Week: Structured-File Ask Continues to Grow
Most of the week revolves around the same request. Publish a structured file for AI to read, and host on your domain.
llms.txt is a cautionary tale here. The file is already there, however Google says that it cannot classify the sites and the data says that the bots did not read it. OKF and ARD are the same application that comes new, their acceptance is not yet confirmed, while the new tools of Bing remain far from the agreement, measuring whether any of these publications actually turn into citations.
The request becomes routine even when there is no payment. Which of these formats benefit from their preservation is still to be determined, and sifting through the winners from the unreadable files is the task at hand this week.
Top Stories of the Week:
Additional resources:
Featured image: PeopleImages/Shutterstock
