Every economy rests on a bargain, and the open web's was unusually elegant. A publisher gave away content for free. A search engine indexed it and sent visitors. Those visitors saw ads or bought things, and that revenue paid for the next article. Nobody signed a contract, but the loop held for two decades and built most of the internet you know.
Agents break the loop at its hinge: the click.
The mechanics of the break
When a user asks an AI assistant a question and gets a synthesized answer drawn from your page, the transaction completes without the visit. The user is satisfied. The model used your work. You received nothing — no pageview, no ad impression, no sale, often not even a visible citation.
This is not hypothetical. On queries where Google shows an AI Overview, clicks to the number-one organic result fall by 58%, up from 34.5% earlier in 2025 (Ahrefs, 2025). The trend is accelerating, and it lands hardest on informational content — the exact material the open web produced most freely, since the overwhelming majority of AI Overviews appear on informational queries.
Who gets hurt
- Ad-funded publishers. Recipe sites, how-to blogs, reference content, news explainers — anything whose model was "rank, get traffic, show ads" faces a structural revenue cut.
- Thin aggregators. Sites that merely restated facts had no moat; the model does that restatement now, for free.
- The commons itself. If producing free content stops paying, less of it gets produced — which starves the very models that depend on it. This is the agent era's tragedy-of-the-commons risk.
Who adapts
Three groups are relatively insulated, and they point to where the economy is heading:
- Brands selling something other than attention. If your content is marketing for a product, being the cited authority is a win even without the click — the user now associates the answer with you.
- Owners of proprietary data. Original research, unique datasets and first-party expertise can't be synthesized from elsewhere. Scarcity has pricing power.
- Direct-audience publishers. Newsletters, memberships and communities that bypass search were never as exposed to the click economy in the first place.
The new deals replacing the old bargain
When an implicit bargain breaks, explicit ones replace it. Four are forming in parallel:
| Model | How value flows |
|---|---|
| Content licensing | AI companies pay publishers for the right to train on or retrieve their archives |
| Paid crawling / retrieval | Per-access or metered fees for AI bots to read content, governed at the protocol level |
| Direct subscriptions | Audiences pay the publisher directly, routing around search entirely |
| Citation as brand equity | Being the named source builds authority and demand even with no click |
Underneath these, the plumbing is being rebuilt too. robots.txt is gaining a Content-Signal layer to separate "you may index me for search" from "you may not train on me." Standards like the Model Context Protocol let sites expose structured, permissioned access to agents. The web is quietly growing a rights-and-payments layer it never had.
What this means for anyone publishing today
The strategic conclusion is not "give up on search." It is "stop being a click-dependent commodity." Concretely:
- Become the cited source, not just a ranked one. Citation is the new impression — it builds brand even without a visit. (See AEO.)
- Build a direct relationship — email, community, product — that doesn't route through an engine's good graces.
- Publish something that can't be synthesized away: original data, real expertise, a distinct point of view.
- Govern your content's AI access deliberately via
robots.txtand Content-Signal, so you decide who reads and on what terms.
The click economy is not dying so much as being unbundled. The brands that thrive will be the ones that figured out what they were really selling — and it was never the pageview.
Frequently asked questions
How do AI agents change the web's business model?
The open web was funded by clicks: search sent visitors, ads or sales paid for content. When an agent reads the page and answers directly, the click and its revenue disappear, breaking the publisher–search bargain and pushing the industry toward licensing, paid retrieval APIs, and per-query compensation.
Will AI kill organic search traffic?
It's sharply reducing it, not eliminating it. Clicks to the top organic result fall ~58% where an AI Overview appears, and that share is growing. Informational content is hit hardest. Publishers must shift toward being the cited source, building direct audiences, and click-independent revenue.
What replaces the click economy for publishers?
Several models in parallel: content licensing deals, paid crawling/retrieval access, direct subscriptions that bypass search, and brand value from being the authoritative cited source. The common thread is reducing dependence on free, ad-funded search traffic.