Google tests a new search economy: AI, ads & publisher payments

Search used to have a fairly simple economic logic. A publisher created useful information, Google helped someone find it, and the user clicked through. That visit gave the publisher a chance to make money through advertising, subscriptions, affiliate revenue, sales or a longer relationship with the reader.

AI search changes that sequence. Google can now answer much more of a question directly, sometimes giving the user enough information before they decide whether to visit the source at all.

That creates a new economic question: If a publisher’s information helps Google produce the answer, what is that contribution worth when the visit itself is no longer guaranteed?

Google is now testing one possible answer. Its early AI Contribution Pilot allows selected publishers to earn money when their content makes a significant contribution to the creation of an AI response in products including Gemini, AI Overviews and AI Mode.

The important change is not yet the size of those payments. We do not have enough public information to judge that. It is the appearance of a second possible economic event: alongside the traditional value of a visit, Google is experimenting with assigning value to the publisher’s contribution to the answer itself.

The old search bargain depended heavily on the visit

Google has never paid ordinary publishers simply because someone clicked an organic result. The value of that click usually appeared on the publisher’s side, where a visit could be monetized or turned into a longer audience relationship.

The basic path was straightforward:

That model was never absolute. A visit did not guarantee revenue, and Google has answered some questions directly for years. But AI answers can move much more of the information-gathering process onto Google itself.

Instead of presenting links and leaving the user to assemble the answer, Google can assemble much of it first. The source still matters, but the visit is no longer necessarily part of every successful search journey.

What Google is testing now

The AI Contribution Pilot is still early and invitation-only. According to reporting by Digiday and Google help material reproduced with that report, participating publisher content can generate earnings when it significantly contributes during the creation of an AI response.

That wording is important because the documentation distinguishes contribution during generation from content that is merely used afterward to verify a fact or provide a link. Contribution is not automatically the same thing as citation.

A page appearing as a source does not, by itself, tell us that the publisher earned money. Nor does the presence or absence of a link reveal how Google valued whatever information helped construct the answer.

The pilot appears to recognize something earlier in the process: the publisher’s information helping make the answer possible. That is different from the traditional referral model, where the economic opportunity for the publisher began when someone arrived on the site.

Does the publisher need the click?

This is also where the claim needs to stay precise.

The public description ties earnings to contribution during the creation of the AI response, and a visit to the publisher’s site is not listed as the qualifying event. That supports the idea that Google is trying to recognize value that can exist before an outbound click happens.

But it does not prove that clicks, engagement or other user behavior can never influence valuation indirectly. We do not have complete public program terms or a verified example showing a specific publisher earning a specific amount from an answer that produced zero visits.

So “Google pays publishers without clicks” is too strong.

The more accurate point is that Google is testing compensation based on contribution to the AI answer while the traditional referral remains a separate possible event. The pilot adds a second economic path; it does not replace the first.

Why the click question matters

This would be mostly theoretical if people behaved exactly the same way when Google showed an AI summary. At least in one large Pew analysis, click behavior differed.

A Pew Research Center analysis published in 2025 examined browsing behavior from 900 U.S. adults and nearly 69,000 Google searches. When an AI summary appeared, users clicked a traditional search result in 8% of visits, compared with 15% when there was no AI summary.

Clicks on links inside the AI summary itself were rarer still: about 1% of visits containing an AI summary.

Those numbers should not be turned into a universal prediction for every query or publisher. The searches that trigger AI summaries may already differ from those that do not, and the study does not prove that AI Overviews caused every difference in behavior.

But it makes the economic question concrete. If more of the informational job is completed on Google, a publisher can contribute something useful without necessarily receiving the same referral opportunity that traditionally followed.

Google says the broader traffic picture is more complicated

Google disputes the idea that AI search can simply be translated into a collapse in publisher traffic.

In an August 2025 post, Google Search head Liz Reid said overall organic click volume from Search had remained relatively stable year over year and argued that the clicks Google sends are becoming “higher quality.” Google defines that partly by whether users remain on the destination instead of quickly returning to search.

That counterpoint matters. A lower click-through rate for some searches can coexist with stable overall traffic if search volume or query patterns change.

But aggregate Google traffic is not the same thing as the economics of an individual publisher. A publisher losing informational visits does not automatically benefit because Google sends more traffic across the web overall, and a “higher quality” visit does not automatically replace the economic value of several visits that never happened.

The AI Contribution Pilot is interesting because it introduces another way for Google to recognize publisher value at a stage before the site visit.

What we still do not know about the money

If this sounds like the beginning of a new publisher revenue model, the biggest problem is straightforward: we do not know what the contribution is worth.

The available material indicates that participants can see earnings information and payment history, but Google has not published a simple rate per answer, contribution or use. There is also no public formula that lets an outside observer verify why one contribution is worth more than another.

We do not know whether valuation depends on frequency of use, the type of information supplied, quality signals, characteristics of the answer or some other combination of factors.

Digiday reported mixed early reactions from publisher sources. Some valued the opportunity to participate and wanted more transparency; others described the financial returns as small or insufficient. Those comments are useful as early signals, but they are not enough to determine whether the program could meaningfully support publishers at scale.

The pilot is also limited and invitation-only. Without broader access, comparable payment data and a transparent valuation method, it is too early to describe contribution payments as a sustainable replacement for traffic-based economics.

Google can monetize the AI surface

This is also why the story matters beyond publishing and SEO.

Google can sell advertising around AI answers. According to Google Ads documentation, ads can appear above and below AI Overviews in supported markets, and Google also supports ads within AI Overviews in certain countries and configurations.

Google has also been testing new advertising formats in AI Mode, including formats designed to place commercial recommendations inside or alongside an AI-assisted search experience.

That creates a potentially important new user journey. A publisher produces information, that information helps Google answer a question, and the user may get enough of the answer without visiting the publisher. Google can still keep the user inside a surface it knows how to monetize.

Under the AI Contribution Pilot, a selected publisher may now also receive some compensation for helping produce that answer.

There is no evidence that a publisher payment is directly funded by the ad shown with a particular AI response, so this should not be described as direct ad-revenue sharing. But the two pieces belong in the same economic picture: Google can create commercial value on the AI search surface while experimenting with a separate mechanism for recognizing publisher contribution.

What this changes – and what it does not

A contribution payment is not economically identical to a website visit. A visit can create advertising revenue, a subscriber, a sale, a repeat reader or a relationship with an audience that exists outside Google.

Publishers can also still receive traffic from AI search. Google continues to include links and source displays that can send users to the open web. So this is not a clean transition from Google sends traffic to Google pays instead.

The pilot is better understood as an additional layer. Selected publishers may still receive visits while also earning something because their information contributed to the answer that came before the visit.

What we cannot yet tell is whether that second form of value could meaningfully compensate for any reduction in the first. The pilot is too limited, the formula too opaque and the public payment data too thin to show that Google is replacing lost publisher traffic with direct compensation.

The conceptual change is still important.

For years, the economic relationship around search was largely organized around referrals: publishers produced the information, Google organized access to it, and the click connected the two business models. AI weakens that connection because useful information can now be synthesized and delivered before the user decides whether another page is necessary.

The AI Contribution Pilot suggests Google is experimenting with recognizing publisher value at that earlier stage.

The click is still part of Google Search. What is new is that the publisher’s contribution may now have economic value before that click happens.