When AI agents start browsing, comparing, and canceling for us, what happens to the sites that used to profit from us doing it ourselves?

This explores what happens to businesses that earn money from human attention or human habits once AI agents do the browsing, comparing and cancelling for people. One thing to know first: the collection has little on news or media publishers directly. What it does have is a sharp picture of the same squeeze in commerce, among travel sites, aggregators and subscription businesses. That is probably the closest early evidence of where publishing is headed.

The clearest damage shows up in businesses that sell placement in front of human eyes. When a shopping or travel agent ranks options against the criteria a user stated, a paid ad slot has nobody to catch. The agent keeps the 'demand signal', meaning the record of what the user actually wants, and never passes it to the aggregator Are AI agents making attention-based ads irrelevant for aggregators?. Tripadvisor's hotel revenue fell 23% in one quarter, and management blamed it on changes in search traffic. A second, less obvious version of the same problem: some business models depend on people forgetting things. Subscriptions that renew by default profit from customers who never cancel. An agent that never forgets to cancel removes that 'inertia subsidy' Will AI agents destroy the negative option billing business model?. So agents threaten businesses that rely on human inattention, not just those that rely on human attention.

The main way businesses are adapting is to compete for the agent itself. One line of research argues that an 'agent attention economy' is replacing the human one. Services will fight to be chosen by agents, which brings new ranking systems, new discovery tools and, predictably, new forms of paid placement aimed at machines Will agents compete for attention just like users do?. Money doesn't disappear in this shift. It moves to whoever runs the agent. One model has an agent's cut per transaction growing from under 1% to about 4%, by adding affiliate fees and eventually booking-site-style commissions How could an AI agent's revenue per transaction grow over time?.

Here is the twist you may not expect: the old conflict of interest comes back inside the agent. An agent paid by merchant referral fees can't be loyal to both the merchant and the user Can an AI agent serve both merchant and user interests fairly?. Agents funded by advertising have a built-in reason to favor sponsored results Do advertising-funded shopping agents favor paid product placements?. One experiment found a surprisingly simple counterweight. Adding a single sentence to an agent's instructions saying the recommendation is for the consumer was enough to restore its resistance to sponsored options Can naming the consumer fix AI agent sponsorship bias?. In other words, which side an agent favors may be set in its instructions, not in its code.

The practical takeaway for a publisher: the revenue doesn't vanish, but it moves upstream to the agent layer. Whether a site shows up in an agent's answer depends on the agent's incentives and on wording in its instructions, not on clicks. For how news or content publishers specifically are responding, such as licensing deals, crawler blocking or pay-per-crawl, the collection doesn't yet have enough material to say.