Google's AI publisher payments could rewrite the SEO playbook

For much of the internet's modern life, publishers have been trying to convince Google to send readers their way. Google's latest experiment suggests the next argument may be about how much they deserve when those readers never arrive at all.

The search giant is testing a programme that compensates publishers whose content makes a significant contribution to AI-generated answers across AI Overviews, AI Mode and Gemini. First reported by Digiday in September, the AI Contribution Pilot potentially introduces a different bargain between platforms and content creators: rather than rewarding visibility, ranking or clicks, Google would pay for information that helps its artificial intelligence produce better answers.

The distinction sounds straightforward. Measuring it is anything but.

According to a September 29 report by The Information, approximately 100 digital publishers are participating in the experiment. Payments vary enormously. Some smaller websites have earned less than $1,000 over several months, with payouts representing under 0.1% of their advertising revenue, while one early participant was reportedly earning at an annual rate exceeding $1 million.

Participating publishers can track earnings through Google Search Console, but the methodology for determining a meaningful contribution remains largely opaque.

The experiment arrives as the traditional search bargain comes under considerable strain. According to Chartbeat data cited by the Reuters Institute, Google organic-search referrals to more than 2,500 websites globally declined 33% between November 2024 and November 2025. Separately, a 2025 Pew Research Center study of US users found that people clicked traditional search links in just 8% of visits featuring an AI summary, compared with 15% when no summary appeared. Links within the AI summaries themselves attracted clicks in only 1% of visits.

For publishers accustomed to exchanging information for traffic, these are uncomfortable numbers.

From being found to being useful

Saurabh Doshi, Co-Founder, OptimizeGEO.ai, believes the experiment changes the incentive around which publishers produce content.

"For the last 20 years, publishers have largely optimised for one thing: getting found, getting ranked and getting the click. This introduces a subtly different incentive, which is whether your content is actually useful enough to shape the AI answer itself," he says.

Doshi draws a parallel with Facebook's Instant Articles, which kept readers within Facebook's environment rather than directing them to publisher websites.

More importantly, he distinguishes between appearing in an AI-generated answer and materially influencing its construction. A publisher might receive a citation without contributing significantly, while original reporting could shape an answer even if nobody clicks through to the underlying article.

That distinction could have significant implications for search engine optimisation, and for the rapidly expanding industry built around generative engine optimisation, or GEO.

Geetanjali Bhattacharji, Co-founder and CEO of Glassbox, a Komerz company, argues that the change goes beyond publishers to brands attempting to establish themselves as authoritative sources.

"For twenty years, content was written to rank. Now it has to be written to be useful, because AI doesn't care how loud you shout, only whether its answer falls apart without you. Every CMO should be asking: if my brand vanished tomorrow, would any AI answer in my category even notice?"

The commercial implications are considerable. Brands have traditionally measured organic search performance through rankings, impressions, website visits and conversions. GEO tools are increasingly tracking brand mentions, citations and share of voice within AI responses.

But these visibility metrics do not necessarily establish whether a piece of information was indispensable to producing an answer.

"Contribution is a much harder thing to measure," Doshi observes, arguing that the next generation of tools will need to look beyond citation tracking to determine how much influence a source actually exercises.

Can you optimise for indispensability?

Pankaj Srivastava, Founder and CEO, UnoSearch, sees the development as an expansion of traditional SEO rather than its replacement.

Publishers, he suggests, will increasingly emphasise originality, factual accuracy, contextual depth and expertise, while also structuring information in ways that AI systems can readily understand.

He anticipates optimisation shifting towards contribution signals, potentially including source credibility, unique insights and structured information.

The obvious question, however, is whether this creates another system for publishers and marketers to manipulate.

Bhattacharji believes an AI contribution optimisation ecosystem will develop faster than traditional SEO, attracting many of the same attempts to game visibility. Yet, she argues, keyword stuffing cannot manufacture genuine indispensability.

Amit Verma, Founder and CEO of Digitup, is more sceptical of the notion that such systems can be easily manipulated.

"To game google is extremely hard," he says, pointing to the importance of demonstrated subject expertise and a consistent history of publishing authoritative information. Although answer and generative engine optimisation are already emerging, he believes genuine experts will remain better positioned to influence responses.

Doshi acknowledges both possibilities. A system rewarding original reporting, fresh data and genuine expertise could strengthen incentives for quality journalism. Equally, once publishers identify what appears to trigger payments, optimisation tactics designed specifically to exploit those signals could follow.

For now, neither the technical criteria nor a common industry standard for measuring contribution has been established.

Two experiments, two ways of paying

Google's latest developments suggest that it is exploring multiple approaches to its relationship with publishers.

On October 15, the company announced that more than 400 publications worldwide now participate in its separate News AI licensing programme. New partners include CNN, Axios, the Financial Times and Forbes. That programme pays publishers to provide content for experimental features such as AI-generated article previews and personalised audio briefings, rather than calculating payments according to individual contributions to AI answers.

The same day, Google announced changes to make links more prominent within AI Overviews and AI Mode, including higher placement of inline links and expanded news carousels, signalling continued efforts to direct users towards original sources.

Meanwhile, publishers are attempting to establish their own standards. The Standards for Publisher Usage Rights coalition, or SPUR, now has more than 80 members across over 20 countries, including The New York Times, AFP and The Economist. Its proposed Content Telemetry Standard seeks to track and report AI systems' use of publisher content, potentially supporting more transparent licensing and compensation.

These developments expose a fundamental question: should platforms independently determine the value of information, or should publishers have a measurable basis on which to negotiate?

Google and multiple publications contacted by exchange4media had not responded to requests for comment at the time of writing.

Google's experiment may ultimately prove to be a modest additional revenue stream rather than a replacement for lost referral traffic. Nevertheless, it introduces a new commercial possibility: content may generate value inside an AI system without generating an equivalent visit to its creator.

SEO taught publishers how to get discovered. The next optimisation economy may revolve around proving that an AI answer would have been poorer without them.

The question is who gets to make that judgement, and what it will be worth.