Reddit has become the single most cited source across ChatGPT, Claude, and Gemini for product and shopping queries — ahead of every retailer, every brand site, every traditional publisher. Review platforms are growing fast as a citation category too. And on brand-specific queries, third-party comparison and roundup content still does a huge amount of the work deciding what a model says about you, even as ChatGPT in particular increasingly links out to brand sites directly once it's formed an opinion. Claude leans more toward synthesising and summarising sources rather than quoting them outright, and Gemini pulls heavily from whatever already ranks well in Google — but all three are drawing from the same underlying pool of affiliate and publisher content to do it.

This is exactly the inventory affiliate programmes already produce: comparison articles, "best of" roundups, honest reviews, Reddit-style community recommendations. The brands showing up in AI answers right now aren't the ones with the best-optimised product pages. They're the ones whose affiliates and publishers happened to write the content the models now treat as evidence.

Content has always been the hardest part of an affiliate programme to justify. A genuinely good comparison guide or in-depth review takes real time and money to produce, the payoff is slow, and it's notoriously hard to attribute cleanly back to revenue. Over the past few years, that's pushed a lot of budget toward publishers where the return is faster and easier to prove — cashback sites, loyalty platforms, and increasingly SaaS and tech partners. Everyone still says they want awareness and editorial reach. Almost nobody could rationalise paying for it the way they could for a performance-driven cashback deal.

That's what makes this moment interesting. If the long-form content publishers were always being paid to produce is the same content AI models are now citing, the case for investing in genuinely useful, well-structured affiliate content just got a lot more reasonable — and not only because it eventually helps your SEO. It's a real argument for funding the slow, expensive, hard-to-attribute content work that's always been the right long-term call but the hardest one to defend in a quarterly budget review.

What I don't have a clean answer to yet — and what came up constantly in Berlin — is what this does to publisher economics from here. If an AI answer increasingly replaces the click that used to justify the commission, does that content get more expensive to commission? Does compensation start moving away from last-click entirely, toward something that credits influence inside an AI answer? Nobody in that room had a confident answer. Neither do I. But it's the right question to be asking before you rebuild a content strategy around it.

When OpenAI started running ads inside ChatGPT earlier this year, organic citations briefly dropped by over 40% as paid placements ate into the space available for unpaid mentions. They've since recovered — but the brands with the strongest existing footprint of third-party coverage recovered fastest and kept their share. The ones with thin or inconsistent coverage lost ground and haven't fully clawed it back.

That's the pattern to take seriously: as AI search monetises and gets more selective, organic citation space becomes scarcer and goes disproportionately to brands that already built an evidence trail. Brands building that affiliate and publisher footprint today are positioning themselves for the next two years of AI-mediated discovery. Brands waiting to "see how this plays out" are choosing to compete for a shrinking amount of unpaid visibility, later, against brands who started now.