Ads for AI agents: Perplexity pulls the plug after less than two weeks

In late July 2026, US magazine Time started selling ads aimed not at human readers, but at AI agents, the ones behind ChatGPT, Perplexity or Gemini, visiting its pages. Less than two weeks later, Perplexity blocked those ads, calling the practice "deceptive". The episode is brief, but it already signals how your content will be read, judged and cited tomorrow.

An ad built to be read by a machine

Time now converts its pages into markdown versions, stripped-down and design-free, to make them easier for AI agents to read. Working with ad tech platform Mobian, the publisher inserts ads formatted as FAQs, written by the brand and labelled "sponsored content". Ally Bank and the Project Management Institute are among the first advertisers. The goal: monetise bot traffic that, on some days, now exceeds human traffic on Time's site.

Perplexity says no

On 11 August 2026, Perplexity confirmed it blocks these ads from its own search index. Its chief communications officer pointed to a reputational and trust-score risk for publishers that rely on "deceptive markdown-style advertising". In plain terms: information written and approved by an advertiser, inserted without clear distinction into the same stream as journalistic facts, blurs the line between editorial content and advertising, for a reader that is, here, an algorithm rather than a person.

An internet splitting in two

This episode illustrates a broader trend. According to Cloudflare, more than half of global web traffic now comes from bots, not human visitors. About 15% of brands already run their own markdown-converted pages to be better read by generative AI, according to Mobian's data. A web built for humans increasingly coexists with a web built for agents, with no shared rules yet in place between publishers and AI platforms.

What this means for your strategy

No standard yet governs how generative AI treats sponsored content slipped into a technical page. Rushing into such an unsettled ad format, risking being labelled deceptive overnight, is not a small decision for a brand. Making sure your site is properly read by AI agents, structured, up to date, with verifiable information, remains legitimate groundwork, useful for humans and machines alike.

The AI view, people first

AI is already changing how your customers find information: a growing share of searches now pass through a conversational agent before reaching a human. That doesn't mean testing every new ad format the day it launches. AI can monitor a brand's visibility in ChatGPT or Perplexity's answers, and flag a change in how it's treated. But deciding where to invest, and where not to venture yet, remains an expert call. AI executes. Expertise decides, and watches.