▶  AI Search · S6E135 · 51:25 · September 7, 2026

ChatGPT Cites Publishers With Deals 48% More with Thomas Peham

Guest: Thomas Peham, CEO & Co-Founder of Otterly.AI
Hosts: Jon Clark & Joe DeVita

Play episode
0:00 51:25
Apple Podcast Spotify Youtube

about this episode

Most AI search advice is theory. Thomas Peham, CEO and co-founder of Otterly.AI, runs a team with two full-time GEO researchers whose only job is to test it. Across millions of citations from ChatGPT, Claude, Gemini, and Google AI Overviews, a consistent picture emerges: corporate and brand websites supply roughly 40–60% of all citations, niche and trade media outperformed mainstream publishers in 16 of 17 US industries studied, and publishers holding licensing deals with OpenAI pick up about 48% more ChatGPT citations than those without. The uncomfortable findings are here too. Schema markup lifted Google organic and AI Overviews visibility but did nothing measurable on ChatGPT. Crawl data shows almost no activity on llms.txt files, while sitemaps get hit constantly. A fictional agency with no clients and no history got recommended by ChatGPT purely by earning mentions in the articles ChatGPT was already citing. And 25% of the URLs ChatGPT cites are dead links. Thomas also shares Otterly.AI’s own Claude numbers, which should make anyone in B2B re-examine their analytics: Claude accounts for 0.1–0.2% of their website traffic, but 10.6% of net-new leads name it as where they first heard about the company.

In This Episode

  • Why hiring dedicated GEO researchers became the marketing strategy
  • Does schema markup actually influence ChatGPT citations?
  • The llms.txt crawl data — and why sitemaps deserve the attention instead
  • Where AI citations come from: corporate sites, niche media, and publisher licensing deals
  • The fake London SEO agency experiment and what it exposes about off-page manipulation
  • Building brand sentiment through reviews, product content, and a central brand entity
  • Why YouTube long-form gets cited regardless of view count
  • What Claude’s citation data reveals about B2B buying you can’t see in analytics

A practitioner’s look at what’s been tested, what held up, and what marketers should stop doing.

Connect with Thomas Peham

Thomas Peham
Thomas on LinkedIn
Thomas on X
Thomas on Instagram

Chapters

  • 0:00

    The 48% Licensing Advantage

  • 1:47

    Meet Thomas Peham of Otterly.AI

  • 4:10

    From Marketing Lead to Startup CEO

  • 6:06

    Why Hire Full-Time GEO Researchers

  • 11:00

    Does Schema Markup Move ChatGPT?

  • 13:16

    Is llms.txt Actually Worth It?

  • 16:35

    Where AI Citations Really Come From

  • 21:18

    The Fake Agency ChatGPT Recommended

  • 24:58

    Brand Sentiment as a Core KPI

  • 28:30

    YouTube’s Rise in AI Citations

  • 32:32

    Reviews, Product Content, Positioning

  • 36:29

    Wikipedia and Dead Citations

  • 41:24

    What Claude Data Says About B2B

  • 46:51

    Rapid Fire: Hires, Advice, Otters

Resources mentioned