An emerging advertising format designed specifically for artificial intelligence agents is facing an early test after Perplexity moved to prevent the ads from influencing its AI systems, raising questions about whether publishers can turn chatbot visibility into a new advertising market.
Brands including Ally Bank and the Project Management Institute have been testing the format on Time’s website through an initiative designed to present advertisements directly to AI crawlers rather than human readers.
The experiment has opened a broader debate across the advertising industry over whether so-called “Agent Ads” can become a viable marketing channel or whether AI companies will block the practice as an attempt to manipulate their search and answer systems.
Perplexity has already taken a hard line.
The AI search company has begun preventing the advertisements from influencing its models, according to Digiday. Perplexity also warned publishers that running what it considers “deceptive advertising” could lead to lower rankings in its search index.
At the center of the dispute is a fundamental question about how information should be presented to AI systems. Time has created stripped-down, text-only versions of some webpages intended specifically for AI agents. The publisher directs AI crawlers toward those versions rather than the pages designed for human visitors.
It then partnered with adtech company Mobian to place “Agent Ads” on the AI-readable pages.
Unlike conventional display advertising, the ads appear as blocks of FAQ-style text containing current facts about the advertiser. The format is intended to provide information that an AI system can read, process, and potentially incorporate into an answer.
The underlying strategy is that if an AI system retrieves information from a respected publisher, the advertiser may have a greater chance of being accurately represented when users later ask a chatbot about the company or its products. That could become more valuable as consumers shift from traditional search engines toward AI assistants for product research, recommendations and factual queries.
Jonah Goodhart, chief executive of Mobian, defended the approach, saying the advertisements are clearly labeled and contain cited, verifiable information that AI systems can decide whether to use.
“This is the opposite of deception,” Goodhart said.
Perplexity disagrees.
Jesse Dwyer, a spokesman for the company, said providing one version of a webpage to human visitors and another to AI systems amounts to “cloaking,” a practice that traditional search engines have historically penalized.
“You can’t just rebrand spam,” Dwyer said. “If the ads are so helpful, why not just put them right on the site for everyone to see?”
The disagreement exposes a potentially important fault line in the emerging AI advertising economy.
Traditional search advertising works largely because users can see the advertisements and distinguish them from organic results. AI systems introduce a different environment. A chatbot may absorb information from a webpage and present it in a synthesized answer without showing users exactly which portions of the source influenced its response.
That creates an incentive for marketers to optimize content specifically for machines, potentially producing a new form of search-engine optimization aimed at AI models rather than people. It also creates a significant governance problem for AI companies. If paid promotional material is deliberately structured to resemble ordinary informational content, platforms must determine whether it should be treated as advertising, useful data, spam, or an attempt to manipulate model outputs.
Perplexity’s response suggests that at least some AI companies intend to apply existing anti-manipulation principles to the new format.
The stakes could rise as AI search becomes more commercially important. If consumers increasingly rely on systems such as ChatGPT, Gemini and Perplexity to research brands and make purchasing decisions, appearing in an AI-generated answer could become as valuable to marketers as appearing prominently in conventional search results. That could create a new competition among publishers, advertisers and AI platforms over who controls the information that models consume.
Time’s position is that the industry needs to develop standards rather than simply reject the concept.
Mark Howard, Time’s chief operating officer, said the publisher had contacted Perplexity and looked forward to “a constructive dialogue about the standards and safeguards that will allow advertising and AI to develop responsibly together.”
The Project Management Institute is taking a similarly experimental approach. Menaka Gopinath, its chief marketing officer, described the organization’s work with Time as part of a “learning journey” aimed at understanding how information about project management, professional skills and career development appears in AI-driven environments.
For advertisers, the immediate value may therefore be less about direct sales and more about learning how AI systems discover, interpret and reproduce information about brands. That distinction could be important because AI advertising remains an immature market, and advertisers do not yet have the same established measurement systems for chatbot visibility that exist for conventional digital advertising.
But the dispute with Perplexity shows that the industry may face an even more basic obstacle: AI platforms control the systems that marketers are trying to influence.
A publisher can create content designed for AI crawlers, but an AI company can decide whether that content enters its search index, influences its models, or appears in its answers. That gives AI platforms significant power over the emerging market.
