Part 1: Platform Fundamentals
What B2B marketers need to know before getting started
ChatGPT Ads are sponsored placements that appear below relevant AI-generated responses in ChatGPT. They are contextually matched in real time based on the content and intent of the active conversation, not on user demographics, browsing history, or keyword bids. The ad unit sits clearly separated from the organic answer and is labeled as sponsored.
This is a fundamentally different channel from paid search or social. There are no match types, no Quality Score, and no audience segments to build. The conversation itself is the targeting signal.
What ChatGPT Ads are not:
- A search replacement. Ads do not appear in the response itself and cannot influence what ChatGPT says.
- Keyword-based. Advertisers provide “context hints” that describe relevant conversation topics, but these are not exact-match keywords and do not guarantee placement in specific conversations.
- Demographically targeted. OpenAI does not share user conversation data with advertisers. No user-level data is available for audience building.
- Performance media at scale (yet). The platform is in active beta. Treat early investment as structured learning, not a scaled demand generation channel.
Ads appear below relevant ChatGPT responses for Free and Go tier users. Each ad unit includes:
| Parameter | Detail |
|---|---|
| Advertiser Name | Brand name as registered in Ads Manager |
| Favicon/Logo | Brand logo pulled from advertiser account |
| Headline (Title) | Primary ad copy, written by advertiser |
| Description (Copy) | Supporting copy describing the offer or product |
| Image Asset | Creative image provided by advertiser |
| Landing page | Destination URL, with UTM parameters persisting on click |
UPDATED – July 2026
Product feed card format: OpenAI has refreshed the product feed card ad unit to surface price and star ratings directly within the ad. This is primarily relevant to catalog or marketplace advertisers. For single-offer B2B SaaS campaigns — the majority of B2B use cases — the standard ad unit structure above remains unchanged.
Ads are currently shown to logged-in adults (18+) on Free and Go ($8/month) tiers. Paid tiers — Plus, Pro, and all Business plans — are ad-free.
| Parameter | Detail |
|---|---|
| Eligible users | Free and Go tier, logged-in adults (18+) |
| Current markets (updated July 2026) | United States, Canada, Australia, New Zealand, United Kingdom, Japan, and South Korea. Self-serve access is live in all seven markets. Mexico and Brazil are confirmed as forthcoming but are not yet available for self-serve activation. |
| Excluded users | Plus, Pro, Enterprise, Business, Education plans; users under 18 |
| Excluded contexts | Sensitive topics (health, mental health, politics); temporary chats; confidential or regulated conversations |
UPDATED – July 2026
The market table has been updated. The prior edition listed US, CA, AU, and NZ as of March 2026. Self-serve is now available in seven markets: US, Canada, Australia, New Zealand, UK, Japan, and South Korea. Mexico and Brazil are confirmed as coming but not yet live. If you are planning campaigns in UK, Japan, or South Korea, account setup can proceed now. For MX and BR, hold account setup until OpenAI confirms availability.
This is the mechanism advertisers have most direct control over, and the one most often misunderstood.
At the ad group level, advertisers provide “context hints” — descriptive phrases that characterize the conversations, topics, or scenarios where their product or service is relevant. OpenAI’s ad system uses these hints as signals, alongside the ad’s landing page URL, headline, and description copy, to match ads to eligible conversations.
NOTE
Context hints are not keywords. They do not guarantee placement in specific conversations and do not use exact-match logic. Think of them as semantic guidance to the auction system, not as bid terms. The richness and specificity of your hint language directly affect how accurately your ads are matched.
The ad selection system evaluates multiple inputs simultaneously:
- Context hints provided by the advertiser at the ad group level
- The ad’s landing page content and URL structure
- The ad’s headline and description copy
- The relevance-weighted second-price auction bid
ChatGPT Ads supports two buying models, each aligned to a distinct campaign objective.
| Parameter | Detail |
|---|---|
| CPC (Cost Per Click) | Clicks objective. Recommended starting max bid: $3-5 USD. Custom max bids available. B2B campaigns may see higher CPCs ($8-15) due to audience depth. |
| oCPC (Cost Per Optimized Click) (added July 2026) | Conversions objective. Billing still runs on valid clicks, but delivery optimizes toward the clicks most likely to convert. Requires conversion bidding enabled and one active standard conversion event via Pixel or Conversions API. Custom conversion events do not qualify. Campaign objective cannot be changed after creation. |
| CPM (Cost Per Thousand) | Reach objective. Default max bid: $60 CPM. Suitable for brand awareness and early-funnel exposure. |
| Auction type | Relevance-weighted second-price auction. Winning ad pays just above the second-highest relevant bid. |
| Minimum budget | No published minimum as of self-serve launch. Recommended test budget: $2,000-5,000 per month for meaningful data collection. |
TIP
Early mover advantage is real. The auction is currently less competitive than it will be as more advertisers enter the platform. Brands building campaign structure, creative learnings, and conversion tracking infrastructure now will have a meaningful cost and knowledge advantage over latecomers.
UPDATED – July 2026
A third buying model is now available. In addition to CPC (Clicks) and CPM (Reach), ChatGPT Ads now supports oCPC — the Conversions objective. You are still billed on valid clicks, but the system optimizes delivery toward the clicks most likely to result in a conversion.
What this requires: Conversion bidding must be enabled, and you need at least one active standard conversion event — set up via Pixel or Conversions API. Custom events do not qualify. One important constraint: once a campaign is created, the objective cannot be changed. Choose before you build.
For B2B: This is directly relevant to decision-stage campaigns. See the B2B Campaign Structure update in Part 3 for how this changes the recommended bid strategy for bottom-funnel ad groups.
Ads Manager Beta currently reports impressions, clicks, spend, click-through rate (CTR), average CPC, average CPM, and conversions. Conversion tracking is available through setup in Ads Manager. UTM parameters can be appended to landing page URLs and persist through ad clicks, enabling standard analytics tools to attribute sessions to ChatGPT Ads traffic.
| Parameter | Detail |
|---|---|
| In-platform metrics | Impressions, clicks, spend, CTR, avg CPC, avg CPM, conversions |
| Conversion tracking | Set up via Ads Manager Beta measurement tools |
| UTM tracking | Static parameters appended to destination URLs; persist on click |
| Advertiser data access | Aggregated performance metrics only. No access to individual conversations, user data, or chat content. |
| Roadmap (expected) | CRM integration (Salesforce, HubSpot, Microsoft Dynamics) planned for later 2026. Closed-loop attribution from ad impression to closed deal. |
| Automatic Advanced Matching (AAM) (added July 2026) | Improves web conversion match rates by using hashed customer data signals. Set up via Ads Manager measurement tools. Recommended for advertisers with existing customer data infrastructure. |
| Mobile Measurement Partners (added July 2026) | AppsFlyer and Adjust are now integrated for app install and in-app event attribution. Relevant only to advertisers with a mobile app conversion path. Web-only B2B campaigns are unaffected. |
UPDATED – July 2026
Two new measurement capabilities are now available.
Automatic Advanced Matching (AAM) improves web conversion match rates using hashed customer data. If you have an existing customer data set or CRM, this is worth enabling — it closes attribution gaps that standard Pixel tracking misses, particularly in low-cookie environments.
Mobile Measurement Partner integrations — AppsFlyer and Adjust are now connected. These cover app install and in-app event attribution for clients with a mobile conversion path. If your B2B product is web-only, this does not affect your setup. If you have a mobile app in the conversion path, configure your MMP integration before launch.
Self-serve access is now available to all advertisers at ads.openai.com. The account structure follows a model familiar from other ad platforms, with one important operational detail:
- Advertiser creates the account. The brand or business registers directly at ads.openai.com. Ownership of the account must sit with the advertiser, not the agency.
- Agency users are invited in. Once the account is active, the advertiser grants access to agency team members who manage campaigns on their behalf. This mirrors the access model used in Google Ads, Meta Business Manager, and LinkedIn Campaign Manager.
CAUTION
Agencies cannot create an OpenAI Ads account on behalf of a client and transfer ownership later. Ensure clients understand they need to initiate registration. For managed clients, coordinate the invite workflow before campaign go-live to avoid access delays.
UPDATED – July 2026
Bulk API now available for agencies managing at scale. OpenAI has added a Bulk API that supports asynchronous, large-volume creation and updates to campaigns, ad groups, and ads. This is relevant if you are managing several client accounts simultaneously or maintaining a large context-hint inventory that requires frequent updates. Standard account management via Ads Manager UI is unchanged — the Bulk API is an additional option, not a replacement for the existing workflow.

