Welcome to Pulse: This week’s updates cover how you measure AI visibility, what to expect from llms.txt, and where the Web Agent is headed next.
Microsoft delivered new AI citation tools, new data landed on llms.txt, Google and two coalitions released agent specifications, and the UK established new ranking rules for Google search.
This is what matters to you and your work.
Bing rolls out AI quote sharing in Webmaster Tools
Microsoft is integrating four new features into the Bing Webmaster Tools AI Performance dashboard. Quote sharing, intentions, topics, and comparison are all in preview at the moment.
Highlights: Citation Share shows the percentage of AI citations your site captures for a given base query, while intents and topics group these queries together to alleviate a data limit in the current dashboard. Compare allows you to overlay a past period on the current period. All four are starting to roll out globally, still in preview.
Why it matters
Citation Share is the first Bing Webmaster Tools metric to show how visible your AI is compared to your competitors, not just whether you’ve been cited. The problem is that it’s all Bing data, covering Copilot and Bing’s own responses, so it doesn’t say anything about Google, where Search Console still offers no citation-style counts.
What SEO Professionals Say
Gianluca Fiorellifounder of ILoveSEO.net, wrote on LinkedIn:
“Bing Webmaster! The Google Search Console we wish we had.”
Read our full coverage: Bing rolls out AI quote sharing in Webmaster Tools
Data from Google and Ahrefs reduces the case for llms.txt
llms.txt received two visits this week. Google’s John Mueller said the file couldn’t help an LLM distinguish one site from another, and new Ahrefs data showed that affected bots barely retrieved it.
Highlights: Speaking on Search Off the Record, Mueller argued that llms.txt cannot differentiate between sites for discovery purposes because the file is self-reported by the very site hoping to be chosen, and he instead referred to plain HTML and internal links. Ahrefs data lands in the same place. Across 137,000 domains, 97% of llms.txt files generated no queries, and crawlers that generate citations, like ChatGPT and Perplexity, accounted for only 1% of the retrieves performed.
Why it matters
The two results point in the same direction. A self-reported file can’t make an LLM choose you, and the bots that generate citations barely pick it up, so don’t expect llms.txt to change your AI search visibility. It still gains a narrow niche with the coding agents and training bots that read it, making it inexpensive to maintain, same conclusion. SE Ranking’s look at 300,000 domains reached months ago.
What SEO Professionals Say
Nat Mileticfounder of Clio Websites, summarized the findings on LinkedIn:
“llms.txt is inexpensive to publish, it’s nice to have it. Don’t expect it to improve AI visibility yet.”
Read our full coverage: Google’s Mueller Says llms.txt Can’t Help LLMs Differentiate Sites And 97% of llms.txt files received no requests, according to Ahrefs data
Google, Microsoft and others release new AI agent specifications
Two agent specs were released within days of each other. Google Cloud released the Open Knowledge Format, and a coalition including Google, Microsoft, GitHub, and Hugging Face followed with Agentic Resource Discovery.
Highlights: The Open Knowledge Format, or OKF, is a markdown format for packaging organizational knowledge, such as datasets, metrics, and runbooks, so that AI agents can read them. Agentic Resource Discovery, or ARD, is a draft specification explaining how agents find and verify tools, skills, and other agents. Both are early, with OKF in version 0.1 and ARD in 0.9.
Why it matters
Neither spec demands anything from you this week. Both repeat the move made by llms.txt, a file structured on your own domain for software to read, and the same unresolved question of adoption hangs over them. Look at what formats are popular before committing to any.
What SEO Professionals Say
Martin Jeffreyfounder and head of strategy at Harton Works, compared ARD to the early days of LinkedIn search:
“It’s the sitemap, reborn for its capabilities rather than its pages.”
Suganthan Mohanadasanco-founder of Snippet Digital, which built two free tools for the format, tempered expectations:
“This is not a magic mushroom and will not increase the visibility of your AI overnight.”
Read our full coverage: Google Cloud announces Open Knowledge format And Google and Microsoft anticipate AI agent discovery specifications
UK orders Google to rank search results fairly
The UK’s Competition and Markets Authority (CMA) has set out new rules for Google Search. Google must rank organic results based on objective criteria and provide notice before any significant changes.
Highlights: The fair ranking requirement covers UK organic results, including AI previews, but not ads, and requires Google to use objective, non-discriminatory criteria, provide advance notice of material changes, and provide a way to raise ranking concerns. A second requirement turns Google’s voluntary UK data portability tool into a legal requirement. Google disputed this assumption, saying its rankings were already fair and transparent.
Why it matters
Notice and complaint requirements are things that could affect daily work, replacing the basic black box update with a warning and a way to respond. A fair ranking also reaches the AI overviews and, like the one from the beginning of June by the CMA unsubscription condition it only applies to the UK, the actual impact depends on how Google chooses to implement it.
What SEO Professionals Say
Laura Iancufounder of Searchpedia, expressed it more bluntly on LinkedIn:
“No more ‘oopsie, we just released another core update.’ »
Chloe Smithhead of strategic SEO at Blue Array, expects negative reactions:
“I expect Google will try to find a way around this.”
Read our full coverage: Google must give notice before significant ranking changes
Theme of the week: the demand for structured files continues to grow
Most of the week revolves around the same request. Publish a structured file for the AI to read and host it on your own domain.
llms.txt is the cautionary tale here. The file already exists, but Google says it can’t tell sites apart and data indicates bots barely read it. OKF and ARD are the same demand coming in recently, their adoption yet to be proven, while Bing’s new tools are on the other side of the market, measuring whether any of these posts actually convert into citations.
The request becomes routine even if the result is not. Which of these formats will win its place is still to be decided, and sorting the winners from the files that no one reads is your job this week.
Top stories of the week:
More resources:
Featured image: PeopleImages/Shutterstock
