Google Is Testing Payments for AI Answers. The Math Is Still a Black Box.
Google is testing a way to pay sites whose content contributes to AI answers. The eligibility rules and economics are still not public.
Google appears to be testing a new kind of content payment: one tied to whether a page materially contributes to a generated AI answer.
That idea is more consequential for marketers than it first sounds. Search Console has historically told teams what their pages received from Google. A reported AI contribution panel would attempt to value what a page supplied inside the answer itself.
The terms remain opaque. Google has not publicly published rates, eligibility criteria, payout timing or a way for most sites to enroll. Content teams should treat this as an early signal about how Google may value grounded answers, not as a revenue line.
The public record starts on June 18, when Markham Erickson, Google’s VP of Government Affairs & Public Policy, wrote that the company was “piloting a new way to partner with websites whose content meaningfully contributes to the freshness and factuality of generative AI responses through the process of grounding.”
Search Engine Journal reported this month that Google confirmed to Digiday it was running an early-stage AI contribution pilot. Screenshots reported by the publication showed a Search Console area for monthly earnings and history. That is the clearest public description of payment, but it comes from reporting and screenshots rather than a public Google program page.
The distinction between being cited and being used
The pilot matters because generative search has created two different questions for a content team.
The first is familiar: did the page rank, get cited or send traffic? The second is newer: did the page help an AI system verify a claim or generate an answer, even if the system did not send a visible click?
The reported help text makes the distinction strict. It says content may confirm facts or appear as a link after an answer is generated without qualifying as AI contribution. The qualifying signal is reported to be a significant contribution during answer generation.
That changes the incentive, at least in theory. A page with clean first-party data, current product facts, an original explanation or a well-supported point of view may help ground an answer even when a basic search-result metric cannot capture the contribution.
It also makes the missing details more important. Google hasn’t explained how it decides that a contribution is significant, how it handles several sources that support the same claim or whether particular publisher categories receive preference. Without that, no marketer can calculate what an editorial investment is worth inside the pilot.
AI-search visibility already differs by audience and prompt. This pilot suggests the underlying use of a source may differ too. The system may use a page as factual support without making it the headline citation a team sees.
What a marketing team can do now
The practical response isn’t to produce content for an unknown payout formula.
It’s to identify the material your site can contribute that a general summary cannot recreate. For a B2B team, that usually means detailed implementation knowledge, independently explainable product limits, benchmark methodology, customer evidence with permission and a clear category perspective.
Those assets already help in ordinary search, sales conversations and AI-assisted research. The value of owned media grows when it helps a buyer frame a decision. The Google pilot gives another reason to make that material factual, fresh and easy to attribute.
A simple editorial audit can help:
- Find pages that make a claim without showing the source or date
- Update product and pricing facts that a buyer would use to compare options
- Put original research methods beside the conclusion
- Preserve the expert explanation that gives a claim context
- Make sponsorship, partnerships and commercial intent clear
Those steps won’t grant access to the pilot. They make the site more useful whether an AI system cites it, reads it for grounding or sends a visitor to it directly.
An early signal, not a publisher deal
Google’s own June announcement speaks broadly about partnerships with websites. Search Engine Roundtable’s report independently described the pilot as a small test. The reported Search Console experience is described as a small, invite-only experiment. There’s no public customer list, no published performance data and no public commitment that the model will expand.
That leaves a meaningful caveat for publishers and brands alike. The pilot could change, remain narrow or never become a broadly available payment program. Google’s separate news partnerships are also not the same thing as this wider web-content experiment.
Still, it exposes a question marketers should watch closely: if a page makes an AI answer fresher or more factual, how does a platform decide who deserves credit?
Google is testing one answer behind closed doors. The teams that benefit most from any future version will already know which parts of their site provide evidence, not just impressions.
Quoted in this story
- Markham Erickson, VP, Government Affairs & Public Policy, Google (source)
Want your perspective in coverage like this? Get quoted.
Sources
- Google Public Policy: Supporting the information ecosystem in the AI era
- Search Engine Journal: Google tests paying publishers for AI answers via Search Console
- Search Engine Roundtable: Google AI Contribution Pilot
This story is part of our running coverage: The State of AI Marketing Budgets →
Get Net Effect, our Monday newsletter.
Net Effect is The State of AI Marketing's Monday recap: how AI is changing marketing, the moves brands are making, and what to do next.
Trouble loading the form? Open the signup form .


