That’s according to a report Tuesday (April 21) from The Information, citing screenshots of the artificial intelligence (AI) company’s ad manager. The news outlet said this confirms its story on the change from last week, which was based on comments from an ad agency executive.
Those new ads, the report added, are in addition to ads that OpenAI introduced in February, which charge advertisers according to how many people view the ads.
OpenAI’s ad manager now lets advertisers choose whether to optimize new campaigns for views or clicks and set the maximum amount they’re willing to pay per click. Advertisers can also provide “context hints” to help inform the types of queries that will generate ads.
The Information noted that OpenAI has set lofty targets for its advertising business, telling investors it has forecast $2.4 billion in ad revenue in 2026 and $11 billion for next year.
However, its earliest ads gave advertisers little more than a glimpse into how they performed, the report added. Marketers say this makes it tough to defend reducing ad spending with the likes of Meta and Google in favor of OpenAI.
Aside from introducing cost-per-click pricing, OpenAI hopes to offer campaigns focused around getting users to take actions such downloading an app or making a purchase, the report said, although the timeline for this rollout remains unclear.
In related news, PYMNTS wrote last month about efforts by the advertising industry to develop infrastructure to deal with AI-related disruptions.
A new initiative known as the Ad Context Protocol is emerging as a standard for AI agents to communicate with publishers, ad platforms and each other. It lets software agents plan, negotiate and activate campaigns via a common machine-readable interface rather than a patchwork of proprietary APIs, the report added, citing coverage in Adweek and Digiday.
“The timing reflects a broader shift underway across digital commerce,” that report said. “As autonomous systems gain the ability to transact on behalf of users or enterprises, industries are moving to define protocol layers before agent activity scales.”
In payments, Visa has debuted developer tools to support agent-driven commerce, championing itself as infrastructure for autonomous transactions, as PYMNTS reported.
Meanwhile, Google has touted its Model Context Protocol servers as a way to standardize the way AI systems retrieve verified data across environments.