Teardown No. 75 · SEO

54 AI Mentions: How Simon Beauloye tracks when AI tools cite his work

Simon Beauloye uses a weekly Claude Code routine to track how often AI tools cite his essays. The scheduled session pulls data from Ahrefs Brand Radar, runs three small scripts in his site's repo and opens one review for him to approve. It writes a cited-by count onto each essay, updates a site-wide mention counter and drafts press-page proposals when a quality publication links to him for the first time. On 2026-10-31 the live counter on his site showed 54 mentions across AI tools. Counts only ever go up.

Simon Beauloye · simonbeauloyeVerified Oct 31, 2026Markdown copy

The problem

He had been wondering how often AI tools were quoting his work in their answers. Readers mentioned seeing his pieces in ChatGPT responses, and he had landed on Perplexity answers that linked back to his site. “But I had no real numbers. Was it ten times a month or two hundred?” He says most authors he talks to are in a similar place. They don’t know how often citations happen, or whether it is worth letting AI agents scrape their content at all.

The blueprint (as he describes it)

  1. Pick a data source. He uses Ahrefs Brand Radar, which watches the major AI chatbots and AI search surfaces and reports when a configured brand or URL appears in their answers. It covers six surfaces: ChatGPT, Perplexity, Gemini, Copilot, Google AI Overviews and Google AI Mode. Claude and Grok are not available yet. The calls reuse a report he has already configured.
  2. Schedule a Claude Code session. The routine is one entry on Claude Code’s /schedule feature, which runs a session on a recurring schedule with the same prompt every time. Once a week the session pulls fresh data from Ahrefs, runs three small scripts in the repo and opens a single review for him.
  3. Guard against double runs. Before doing any work, the routine checks whether it already ran in the last 24 hours. If it did, it exits. This stops a scheduled run and a manual run from doubling up on the same data and burning the Ahrefs quota.
  4. Split the API calls. Each run makes seven calls to Ahrefs: three for per-article citations, three for the site-wide counter and one for press proposals. Brand Radar refuses to mix Google data sources with each other or with the chatbot sources in a single call. The routine queries each one separately and stitches the results back together.
  5. Hand the writing to scripts. One script walks the citations into the right article. One updates the file behind the site-wide counter. One drafts the press proposals. He calls the Claude Code routine “just orchestration”.
  6. Put the numbers on the page. Each essay ends with a cited-by line: one icon per AI tool and one count per icon. A tool that has never cited the article does not appear. Hovering an icon shows the prompt that triggered the citation when one is available. A second line on the home page, about page and press section shows the cumulative count of AI mentions of his name.
  7. Publish the same data as structured data. Each per-article citation is added to the article’s structured-data block as an entry naming the AI tool. The site-wide totals go into the home page’s structured profile block as a per-tool counter. The visible HTML and the structured data carry the same numbers.
  8. Never let a count go down. Once an article gains an AI citation it stays in the records, even if the response disappears from Ahrefs the following week. In his words: “Removing visible social proof retroactively because of an API blip would be misleading.”
  9. Review everything. The two derived signals could go straight to the live site, but they don’t. “Same review surface, single audit trail.” First-time links from quality publications arrive as a draft file in the press folder via a GitHub PR. He reviews the proposal, adds the outlet’s logo, tightens the blurb and decides whether to feature it. He says the weekly review is also a way to catch trouble early.

Stated results

  • Site-wide counter as shown on 2026-10-31: “Cited by AI 54 mentions across AI tools”. Gemini 16, Perplexity 13, ChatGPT 12, Google AI 10, Copilot 3.
  • The essay’s own cited-by line: ChatGPT 4, Perplexity 4, Gemini 3, Copilot 2, Google AI 1.
  • Effect on how often he gets cited, or on traffic: None stated.
  • Planned next: a per-article delta in the weekly review, then a build-time check that flags when the site-wide counter hasn’t refreshed in 14 days.

Caveats

  • The data source is a paid Ahrefs product. He writes that the Ahrefs piece “costs whatever your Brand Radar plan costs.” His links to Ahrefs carry referral tracking parameters. The source does not say whether he has any arrangement with Ahrefs.
  • He offers to share the specs with readers who reach out. No product, course or kit is sold on the page.
  • The site-wide counter tracks brand-name mentions, and the per-article lines track URL citations. He says the two counts don’t add up.
  • Counts are cumulative by design and never reduced, so the 54 figure is a running total rather than a current snapshot.
  • He says he does not know whether the structured data changes how often ChatGPT or Perplexity feature his content. He publishes it because it “costs nothing to generate.”
  • Claude and Grok are not covered by Brand Radar yet.
  • Handles: LinkedIn linkedin.com/in/simonbeauloye, X @simonbeauloye, Instagram @simonbeauloye (verified: linked from his own site).

FAQ

What does the 54 figure count?

It is the cumulative number of times AI tools have mentioned his name in their answers, as shown on his site on 2026-10-31. It is a different signal from the per-article URL citations, and the two counts don’t add up.

Which AI tools does he track?

Six surfaces through Ahrefs Brand Radar: ChatGPT, Perplexity, Gemini, Copilot, Google AI Overviews and Google AI Mode. Claude and Grok are not available yet.

Does publishing citations as structured data get him cited more?

He says he doesn’t know. He publishes it because it costs nothing to generate.

Does anything go live without his approval?

No. All three signals sit behind the same weekly review, and press proposals stay as drafts until he does the editorial work.

Sources

Want agents like these working your distribution? Get It Built. Get new teardowns weekly: subscribe (double opt-in).

Get one teardown a week.

Same format, a new operator every issue. Double opt-in.