OpenAI Adds Pixel + Conversions API to ChatGPT Ads Manager

OpenAI opened ChatGPT Ads Manager to all U.S. businesses on May 5, ending the 600-advertiser pilot and adding the two missing pieces our earlier reporting flagged: a Conversions API and pixel-based tracking. The platform also expanded the ad pilot into five new countries (United Kingdom, Mexico, Japan, Brazil, South Korea) and named the agency partner network (Dentsu, Omnicom, Publicis, WPP) plus tech partners (Adobe, Criteo, Kargo, Pacvue, StackAdapt).

For B2B teams that watched the April CPC pivot from the sidelines, May 5 is the more consequential update. The single biggest objection to running ChatGPT ads in a serious paid-search budget has been the absence of post-click attribution. With pixel tracking and a Conversions API, B2B marketers can now measure what happens after a ChatGPT click (lead form fill, demo request, trial signup) with the same plumbing they already use for Google Ads and Meta. The recommended CPC range stays at $3 to $5 per click.

The agency partner list is the second-order signal worth reading. Dentsu, Omnicom, Publicis, and WPP only show up on a vendor’s partner page when the holding companies have committed pilot dollars. Their presence means a meaningful share of next-cycle B2B media plans now has ChatGPT as a line item, which raises the auction floor and squeezes the early-mover CPC advantage faster than B2B teams are typically priced for.

What Changed Since Our April Reporting

The two existing IVRIS pieces on ChatGPT ads cover the February launch (ChatGPT ads went live with a cost-per-engagement model) and the April CPC pivot (CPMs collapsed from $60 to $25, OpenAI pivoted to $3-$5 CPC, and the $50K minimum opened mid-market access). The April piece flagged “the measurement problem nobody has solved” as the remaining barrier for B2B teams running ChatGPT in a serious paid-search rotation. May 5 closed that gap.

Pixel tracking lets advertisers measure on-site events after a user clicks a ChatGPT ad. The Conversions API lets the advertiser send server-side conversion events back to ChatGPT for optimization, the same architecture Meta and Google rolled out years ago for iOS-and-cookie-impacted attribution. Together, the two pieces let a B2B advertiser optimize ChatGPT campaigns toward lead-form fills or pipeline events instead of clicks alone. CPA bidding (cost-per-action) is the next promised step; Digiday reported it as in development but not yet live.

The B2B Auction Math Just Got Steeper

Self-serve plus four agency holding companies plus a Conversions API is the standard pattern for a paid-search auction to mature. Two things happen next, on a predictable timeline: pilot brands publish case studies (six to eight weeks out), and the holding companies start moving budget from underperforming legacy channels into ChatGPT (ten to fourteen weeks out). Both raise the floor CPC fast.

For B2B specifically, the auction implications are sharper than for direct-to-consumer. B2B has fewer searchers, more concentrated keyword sets, and higher-LTV conversion events. When a holding company moves $5 million of a Fortune-500 B2B client’s budget onto ChatGPT, the entire B2B SaaS auction in that vertical reprices. The window for “test ChatGPT ads at sub-$5 CPC before the auction crowds” is measurable in months, not quarters.

The split-screen our April piece noted is intact: ChatGPT Plus, Pro, Business, and Enterprise users still see zero ads. The ad surface only reaches Free and Go ($8/month) tier users. For B2B advertisers, that means the paid track must run alongside an earned AEO and citation strategy aimed at the paid-tier audience that the ad will never reach.

What About Perplexity and Google AI Mode

The same week OpenAI opened the Ads Manager, the competitive map clarified. Perplexity reaffirmed its no-ads stance and pitched “curiosity users” as the value proposition. Google ran a parallel track on the organic side with AI Overview’s Further Exploration update and on the paid side with AI Mode ads. Three different AI-search surfaces, three different monetization stances, one B2B marketing team trying to allocate against all of them.

Our read: ChatGPT Ads Manager is now the only AI-search ad surface with self-serve + CPC + Conversions API + named agency partners. Google AI Mode ads are the closest competitor but route through Google Ads, which most B2B teams already run. The competitive question for the next budget cycle is not whether to test ChatGPT, but how much budget to move from Google Search to ChatGPT before the auction stabilizes.

What B2B Teams Should Do in the Next 30 Days

1. Install the pixel and configure the Conversions API. Both ship with developer documentation comparable to Meta’s. The work is two to four engineering days for most B2B teams. Until the pixel is firing, every ChatGPT click is unattributable, and the auction will move past the test window before measurement catches up.
2. Run a $5K-$15K test budget across two campaign types. One brand-defense campaign on the company name and category, one prospecting campaign on a high-LTV use case. Use the pixel to measure lead-form fills, demo requests, or trial signups. Compare CPL against current Google Search CPL on the same keyword set. The data will be noisy at first; the directional read is whether ChatGPT lead quality matches Google’s.
3. Map the agency conversation. If your media plan runs through Dentsu, Omnicom, Publicis, or WPP, ask which ChatGPT pilots the agency is currently running and what early CPA data exists. Holding-company-side data is typically two to three weeks ahead of public case studies and lets B2B teams size the auction shift before the public CPC moves.

The mid-market and enterprise B2B teams that close the Conversions API integration in May will be the ones with credible Q3 ChatGPT-vs-Google attribution data when the auction starts repricing. The ones that wait for the case studies will be paying the post-stabilization CPC.

Frequently Asked Questions