AI Citation Share Ships, New Data Doubts LLMS.txt – SEO Pulse

Search Engine Journal’s weekly SEO Pulse digest covers a busy stretch: new Bing Webmaster Tools features for measuring AI citations, fresh evidence that llms.txt is largely ignored by the bots that matter, two nascent agent-discovery specs, and the UK’s CMA imposing fair-ranking rules on Google. The through-line is the growing stack of structured files sites are being asked to publish for AI consumption — and the widening gap between publishing effort and proven payoff.

What’s actually new

Bing Webmaster Tools added four preview features to its AI Performance dashboard: Citation Share (your site’s percentage of AI citations for a query versus competitors), Intents, Topics, and Compare. This is Bing-only data covering Copilot and Bing answers — Google Search Console still offers nothing comparable. Separately, Google’s John Mueller publicly argued that llms.txt can’t help LLMs differentiate between sites because the content is self-reported, and Ahrefs data across 137,000 domains showed 97% of llms.txt files received zero requests, with citation-generating bots like ChatGPT and Perplexity accounting for just 1% of the fetches that did occur. Meanwhile, two new structured-file specs appeared: Google Cloud’s Open Knowledge Format (OKF, v0.1) for packaging organizational knowledge, and Agentic Resource Discovery (ARD, v0.9), backed by Google, Microsoft, GitHub, and Hugging Face, for agent-to-agent tool and skill discovery. Finally, the UK CMA ordered Google to rank organic results — including AI Overviews — using objective criteria, give advance notice before significant changes, and provide a complaints route.

What it means for your config

This roundup doesn’t touch build tooling or application config directly, but if your site maintains an llms.txt file, the data here is worth internalizing: the file is cheap to keep, but the evidence says it isn’t moving the needle on AI citation visibility today. For teams evaluating OKF or ARD, both specs are too early to warrant production adoption — OKF is at 0.1 and ARD at 0.9 — and neither has demonstrated meaningful crawler uptake yet. There’s no config migration or breaking change to act on. The Bing Citation Share metric is purely an analytics dashboard addition; it doesn’t require any site-side file or markup changes. The UK fair-ranking requirement could eventually affect how Google surfaces content in AI Overviews, but its practical impact depends entirely on implementation details that don’t exist yet.

Recommended next step

If you already have an llms.txt file, leave it — the maintenance cost is near zero. Don’t invest time creating OKF or ARD files until adoption signals emerge from the bots you actually care about. The most actionable item here is Bing’s Citation Share: if you’re not already checking the AI Performance section in Bing Webmaster Tools, it’s now the only place offering competitive AI citation data in a first-party dashboard. Bookmark it, set a reminder to revisit once it exits preview, and read the full digest for the quotes and context that didn’t fit here.

---

Read the full announcement on Search Engine Journal → AI Citation Share Ships, New Data Doubts LLMS.txt – SEO Pulse