OpenAI is building an internal tool to measure whether advertisements served inside ChatGPT convert users, according to a Digiday report. The development amounts to the company laying the measurement foundation for a planned in-chatbot ad business, and signals OpenAI is past the planning stage on at least one piece of an advertising product it has hinted at for months.

Tracking ad conversions is table stakes for any large ad-supported platform. Google, Meta, Amazon, TikTok and Microsoft all operate attribution systems their advertisers rely on. OpenAI is now building the equivalent for a surface that breaks most of the assumptions those systems were built on.

What does the tool actually do?

Per the report, the system is designed to:

• Log a user's exposure to a sponsored message inside a ChatGPT session
• Record follow-up actions taken by that same user, whether inside the same thread or after leaving it
• Match the two events to produce a conversion signal advertisers can act on

OpenAI has not commented on the Digiday report or confirmed the tool's status.

Why doesn't standard ad measurement work in a chatbot?

Conventional digital attribution relies on three primitives: cookies set on a browser, pixels that fire when a page loads, and URL redirects that carry a click identifier across the open web. ChatGPT offers none of these.

Conversations happen inside OpenAI's own interface. There is no external page for an advertiser to drop a pixel on. Any tracking has to happen server-side, inside OpenAI's infrastructure, with session identifiers only OpenAI controls. That gap is what the new tool is built to close.

Why would advertisers buy ChatGPT ad inventory?

Two reasons stand out.

First, return-on-ad-spend math is the reason most large advertisers buy display or paid-search inventory at all. Without a credible conversion signal, the buy becomes brand-awareness spend, which carries a smaller budget pool.

Second, ChatGPT's reach represents a surface large enough that consumer-goods advertisers will at least test it. They will not commit meaningful budgets without a measurement layer behind it.

How does this compare to peers?

The build is being tracked closely by:

• The Trade Desk, which would face an ad-network competitor with a first-party data advantage
• Google, whose search-ad dominance is the obvious long-term target
• Amazon, currently the third-largest digital ad platform by revenue
• Independent adtech vendors that may build complementary measurement if OpenAI exposes APIs

What hasn't OpenAI disclosed?

The Digiday report did not include:

• A ship date for the tool
• Names of advertisers currently testing it
• The pricing model OpenAI plans (CPC, CPM or hybrid)
• Whether ChatGPT users will see explicit "sponsored" labels on ad responses
• How the company will handle existing Plus and Team subscribers who already paid for an ad-free tier

Those gaps are unusually wide for a product reported to be close to launch. They suggest the ad business is being assembled in pieces rather than rolled out as a finished package.

What does this cost — and what doesn't it?

OpenAI has not disclosed what the tracking layer will charge advertisers, whether access will be bundled with broader ChatGPT Enterprise deals, or whether the data will feed back into model training. Each of those questions carries competitive consequences that go beyond ad operations into AI product strategy itself.

The bigger commercial question for OpenAI is how it monetizes a conversational product without eroding user trust in the model's answers. A measurement tool addresses one half: proving value to advertisers. The other half — proving non-manipulation to the people typing the prompts — is not yet in the public reporting.