In short: if AI search takes a large share of informational referrals, internet revenue will not disappear. It will shift away from pages that monetized the visit itself and toward businesses that control transactions, direct customer relationships, proprietary data, tools or infrastructure. The weakest position belongs to sites whose main product was an advertising pageview on generic information.
This does not mean information sites stop making money. The point at which value is captured changes. If a user gets an explanation directly inside Google, Gemini or ChatGPT, the publisher cannot show that user an ad on the article page. But the same user may still buy a product, book a hotel, purchase insurance, subscribe to a service, use a professional database or hire a provider.
From the information click to the completed action
The old search economy worked because a webpage usually sat between a question and its answer. Users searched, clicked and publishers monetized attention through advertising, affiliate links, sales or subscriptions.
If AI answers the question inside the interface, the first part of that chain shrinks. The Reuters Institute reports that publishers expect search traffic to decline by roughly 43% over three years. At the same time, Google is developing agentic commerce and tools designed to move users closer to completed purchases.
The first winners are the AI platforms
The clearest beneficiary is the company controlling the place where the user asks the question. In traditional search, Google monetized ads around results while much of the interaction happened on other websites. In a generative interface, users can stay longer inside the platform, refine questions, compare options, request recommendations and increasingly move toward action without leaving.
If this becomes mainstream, the platform gains power across several stages: it knows the initial intent, controls the AI answer, decides which sources and brands appear, and can participate in the commercial transaction that follows.
Stores and services are stronger than advice pages
AI can explain how to choose a washing machine, but somebody still has to sell it. It can compare insurance policies, but a policy must still be issued. It can describe an engine repair, but a garage still has to perform the work.
Websites tied to real transactions are therefore less dependent on informational referrals alone. Their risk shifts toward platform dependency and margin pressure rather than simple disappearance.
An audience you do not need to reacquire becomes more valuable
For information businesses, one of the most valuable assets may be the habit of returning. Reuters Institute reports that subscriptions and membership remain among the main monetization priorities for commercial publishers.
If a reader arrives only through a generic query, the relationship lasts minutes. If they know the brand, open an app, read a newsletter or pay for access, Google no longer controls every interaction. Brand therefore becomes an economic asset rather than a cosmetic marketing layer.
Proprietary data can be monetized without a huge audience
A stronger model emerges when a site owns structured information rather than only prose: price histories, legal databases, sports records, long-term measurements, rankings with transparent methodology or professional company data.
Such assets can be sold through subscriptions, APIs, professional access, enterprise plans or AI licensing. In this model, the information site begins to resemble SaaS: articles explain and attract, while the real value sits in the underlying resource.
The machine itself can become a customer
On September 30, 2026, Cloudflare launched the beta version of Pay Per Use, allowing content owners to accept prices for specific machine use of their material.
Large publishers already have direct licensing deals with AI companies. Reuters Institute reports that 37% of publishing executives see platform payments through licensing or similar mechanisms as an important growth opportunity, although only around 20% expect these revenues to become substantial.
Infrastructure owners can earn regardless of who is reading
AI search requires cloud computing, payments, identity, APIs, databases, bot protection, analytics and commerce protocols. If an agent finds and buys a product, somebody processes the payment. If a model accesses a private database, an API and permissions system are required.
This means a larger share of web revenue may flow through payment networks, cloud providers, marketplaces, ad platforms and other infrastructure layers.
| Business model | How it earns | What AI does to informational clicks | Position |
|---|---|---|---|
| Generic information site with display ads | Pageviews | Removes some referrals | Weak |
| Basic affiliate site | Merchant referral | Moves comparison into AI | Weak/medium |
| Subscription author-led resource | Direct audience payments | Can replace only some answers | Strong |
| E-commerce store | Product sales | Helps users choose, but purchase still happens | Strong |
| Local service | Physical service | Can generate a lead, not perform the work | Strong |
| Database / professional tool | Access, API, subscription | Can use data but cannot easily replace the product | Very strong |
| AI platform / search engine | Ads, subscriptions, commerce | Controls the interface | Very strong |
| Payment and technical infrastructure | Fees and services | Earns from executing the action | Very strong |
Affiliate economics may change sharply
Affiliate sites historically monetized the middle step in a purchase decision. A user searched for a product, opened a comparison and clicked through to a retailer. AI can absorb that middle layer by asking about budget and requirements, comparing products and recommending options directly.
Affiliate commissions may survive, but the recipient could change. Platforms that actually shape the decision may demand a larger share of the value that once went to intermediary sites.
Creator businesses also increase the pressure
The information web is competing not only with AI but also with creators who have recognizable voices and multiple monetization channels: advertising, subscriptions, sponsorships, paid communities and their own products. Such businesses depend less on any single search referral.
Where the money may move
If informational referrals fall sharply, the likely direction is from accidental attention toward direct audience relationships; from generic information toward proprietary data; from reading toward action; from pageview advertising toward subscriptions, transactions and licensing; and from the webpage itself toward services, databases, tools and brands.
Advertising will not necessarily disappear, but more of the commercial decision-making may happen inside the AI platforms that control the interface.
The most valuable asset may be control of the next step
In an AI-first internet, it may be more useful to ask what happens after the information is delivered and who controls that action. If the next step is a purchase, sellers and transaction platforms are strong. If the user needs professional analysis, databases and specialist services capture value. If people want a recurring viewpoint, authors can monetize direct relationships through subscriptions.
The weakest position belongs to businesses that control nothing after the answer and own only a page users once had to open to read a few paragraphs. Falling information clicks should therefore be understood not only as an SEO crisis, but as a broader redistribution of internet income.
