OpenAI's ChatGPT Business Hub gives publishers first-party AI data — with a legal catch
OpenAI's ChatGPT Business Hub exposes mentions, clicks and discovery topics for the first time, but its terms may require publishers to waive certain legal claims to get the data.
Wire notes
- OpenAI launched the ChatGPT Business Hub last week, reporting mentions, profile views, clicks and discovery topics for verified business accounts.
- OpenAI's terms require businesses not to pursue certain legal claims, including disputes over content published before signup; a spokesperson said the terms were not intended to cover this scenario.
- Half of the half dozen large publishers Digiday spoke with were setting up business accounts.
- Lily Ray recorded 4 clicks from 25 mentions on her consultancy's dashboard.
- The hub's data is aggregated, with no prompts, per-article breakdown or competitor benchmarks.
OpenAI's new ChatGPT Business Hub gives publishers their first first-party view of how they appear in ChatGPT — but signing up may require waiving certain legal claims against the company. The dashboard, launched last week for businesses that claim and verify their accounts, reports mentions, profile views, website clicks and the topics driving discovery. Publishers previously had to estimate this activity with third-party tools (analytics firms that run prompts at scale) or infer it from referral traffic after the fact.
"Until now, everything publishers knew about their presence in ChatGPT came from third-party tools running prompts at scale and estimating, or from referral traffic after the fact. This is a very early, very limited Search Console moment for ChatGPT," said John Shehata, CEO and founder of publisher SEO analytics firm NewzDash.
How are publishers reacting?
Publishing executives are split. Of the half dozen large publishers Digiday spoke with, half were in the process of setting up business accounts.
One industry executive, who requested anonymity, flagged a legal risk in the fine print. OpenAI's terms require businesses to agree not to pursue certain legal claims against the company, including disputes involving content published on their websites before they signed up. For publishers already grappling with how to protect and monetize content in the AI era, trading legal rights for visibility data may give some pause. An OpenAI spokesperson said the terms "were not intended to cover this scenario."
Can the data support GEO businesses?
Two publishing executives said they were interested in using Business Hub insights to support their GEO (generative engine optimization, the practice of optimizing content to appear in AI answers) offerings, because AI attribution and measurement remain a major hurdle.
"These insights potentially provide a really valuable piece of the puzzle that should validate our own success criteria around visibility and citations," said Carly Steven, Daily Mail director of SEO and editorial ecommerce.
Shehata offered a caveat: a publisher's own citation count does not prove a client's campaign delivered results. "With authorized access to a client's data, [publishers can] finally prove that recommendations move real platform metrics. That closes the measurement loop missing from every GEO pitch. But a publisher's own citation count is not proof of a client's GEO return. A brand campaign still needs evidence at the brand level. The commercial opportunity is selling better evidence and better decisions, not citation counts," he said.
One executive was blunter: because the hub tracks only mentions of the business owner, it misses the scenario where a publisher's sponsored content about a brand's product gets cited — the mention of the product never reaches the publisher. "At this stage, I see brand mentions in answers akin to billboards. It makes sense for Nike to have a billboard but not [a publisher]," the exec said.
What are the measurement limits?
Lily Ray, vp of SEO strategy and research at Amsive, noted the hub tracks mentions of a business website rather than its brand name, a "limited view" that also makes click-through rates appear disproportionately high: her consultancy's dashboard recorded four clicks from just 25 mentions. She called the first-party data "another data point for a very convoluted channel" that "is not going to replace any of the tools I'm using."
Shehata recommends three steps: set a baseline for mentions, clicks and discovery topics; use discovery topics — "the most valuable data for audience teams" — to steer editorial experiments; and separate visibility, audience and business outcomes. "If reported visibility grows while referrals stay flat, that raises a hard question about value the dashboard alone won't answer," he said. The data is aggregated, with no prompts, no per-article breakdown and no competitor benchmarks.
Barry Adams, founder of Polemic Digital, said the lack of context makes reports hard to use, but pointed to Google's Search Console, which gained useful tracking over time. The open question is whether OpenAI follows the same path — and whether publishers will accept the terms before it does.
via seroundtable.com (Original)
More from Amara Osei
Show full bio
News editor covering people, launches and day-to-day practice in marketing, advertising, marketing technology, media, SEO and search at Marketing Herald.
354 articles