--- title: "ChatGPT ads are open for business. Pharma mostly isn't on the list." date: "2026-10-12" summary: "What pharma can do with ChatGPT ads under OpenAI's policy: no Rx promotion, a narrow unbranded lane, no ads in personal health chats, and how to measure it." tags: [paid-search, aeo] --- Every brand team I talk to is asking whether they should buy ChatGPT ads. For a prescription brand, the answer right now is no, because you can't. There's a narrow unbranded lane that may be worth testing, and a bigger opportunity that doesn't involve buying media at all. ## Where things stand OpenAI started testing ads in the US on February 9, 2026, on the Free and Go plans. Plus, Pro, Business, Enterprise, and Edu accounts don't get ads, and neither do Temporary Chats or users identified as under 18. Ads appear below a response and are labeled as sponsored. In May, OpenAI launched a beta self-serve Ads Manager, added CPC bidding on top of CPM, and announced a Conversions API and pixel. On August 31 it reported a $1 billion annualized run rate. As of September 23, it says ads are available in more than 60 countries. Ads are matched mostly on the current conversation, plus general location, the advertiser's targeting choices, and past chats if the user has personalization turned on. OpenAI hasn't published a full list of targeting options, and I've seen nothing about HCP targeting. ## What the health policy says Read the [ad policies](https://openai.com/policies/ad-policies) yourself; they've been revised six times since March. Here's what matters for pharma. Regulated medical products and treatment claims are restricted, and prescription drugs aren't on the allowed list, so branded Rx is out. Health ads are US only and approved one by one. Allowed categories include devices and apps, diagnostics, insurance, hospitals, and "disease awareness campaigns that do not advocate for public policy." Unapproved products and "experimental or investigational therapies" are banned. The policy doesn't mention trial recruitment, so ask OpenAI before you assume either way. Advertiser eligibility is a separate check. Companies whose primary business falls in a restricted category, with healthcare named as an example, "may be ineligible." A manufacturer may not get approved even for unbranded work. Placement is its own filter. No ads appear in "mental and personal health conversations." General medical-advice chats haven't been blocked by default since April. But "I was just diagnosed with..." counts as personal, and that's where most patient intent sits. Even an approved advertiser only shows up next to general health conversations, and OpenAI's classifier decides which ones qualify. ## What's realistically possible Unbranded disease awareness is the one path the policy names. That means symptom literacy or what a specialist workup involves, linked to an unbranded site. It still needs MLR review, and OpenAI reviews the landing page too. Diagnostics are a named category. That matters if you or a partner sells a test. HCPs: mostly no. There's no profession targeting, and clinicians on enterprise or paid accounts never see ads. Copay and support programs: assume no. A copay card names the product, which makes it promotion. ## Questions for MLR Adjacency. Your ad sits under an answer you didn't write and can't preview. If that answer names a competitor's drug, does your ad imply a comparison? Decide, and document it. Conversations. Users can "Ask ChatGPT" about an ad, and they can message an advertiser through it. Those messages need adverse-event and medical-information routing. Pixels on health pages. Reporting is aggregated, but OpenAI's pixel on a disease site is still a third-party tag on a health property. Have privacy counsel review it. Policy churn. Check the changelog before every flight. FDA is also moving to tighten DTC drug advertising rules, so keep any test unbranded and boring. ## How you'd know it worked You get CPM or CPC delivery and aggregated conversions. Pick one action on the unbranded site as the conversion, such as a discussion-guide download. Set a cost ceiling and compare against your search baseline. If you can't define the action before launch, you aren't ready to buy. ## Paid placement won't get you into the answer OpenAI says ads "do not influence ChatGPT's answers." Take that at face value: an ad does nothing for whether you're cited in the answer above it. Patients read that answer first, and many of their decisions happen in personal health chats that never carry ads. In our audits, the sources behind that answer are often advocacy groups, reference publishers, and competitors' unbranded content. Getting your MLR-approved language into that source layer will do more for you than an ad underneath it. ## Takeaway For Rx brands, ChatGPT ads are mostly closed. If OpenAI approves you, test unbranded disease awareness and measure it like any new channel. Spend most of your effort on the answer itself.