The ChatGPT Ads Manager is where you set up, run, and measure ads inside ChatGPT. It's OpenAI's self-serve interface at ads.openai.com, which opened in the US in May 2026 and has since expanded to more markets. This walkthrough covers what the ChatGPT Ads Manager looks like today: account setup, campaign types, targeting, creative, budgets, and reporting. Because OpenAI is shipping features quickly, treat the specifics as a snapshot we update over time.
New to the channel? Start with our overview of how ChatGPT ads work in 2026. This piece assumes you already know the basics and want to see the platform up close.
Getting into the Ads Manager
Access begins at ads.openai.com. You create an advertiser account, add billing, and verify your business. The self-serve Ads Manager is available in seven markets as of mid-2026 (the US, UK, Canada, Australia, New Zealand, Japan, and South Korea) and runs with no minimum spend; if it isn't open in your market yet, an agency in a live market can run campaigns on your behalf.
Verification is the step that trips people up. Expect to confirm your business details and a payment method before you can spend, and budget a little lead time while the account is reviewed. Getting this done early means you can launch the moment your creative and landing pages are ready, instead of waiting on approvals when the campaign is due to go live.
The dashboard: what you see
The layout will feel familiar to anyone who has used Google Ads or Meta Ads Manager. You get an account level, a campaign level, and an ad level, with a reporting view across all three. The mental model is the same as other auction platforms: campaigns hold your objective and budget, ad groups (or their equivalent) hold targeting, and ads hold the creative that appears under answers.
Campaign types and objectives
Objectives map to where you want the placement to take someone. Early on the choices are lean, which is normal for a young platform. Expect to pick between awareness-style goals and click or traffic goals, with more conversion-focused objectives arriving as measurement matures.
- Awareness or reach: buy impressions in front of high-intent users on a CPM basis.
- Traffic or clicks: pay per click (CPC) to drive users to your destination.
- Conversion-oriented objectives: expanding as OpenAI builds out attribution.
Targeting options
This is the part that differs most from legacy platforms. ChatGPT's value is intent: people ask detailed questions, and the ad is matched to the context of the conversation rather than to a cookie profile. In practice, targeting leans on query and topic relevance more than on granular demographic segments. OpenAI has been explicit that advertisers don't get access to individual conversations, so you're targeting context and category, not reading anyone's chats.
Practically, that means your job is to be genuinely relevant to a question. The more precisely your product answers a real query, the better the placement performs. Broad, spray-and-pray targeting works worse here than on display.
Creative and ad formats
In the manager you build the Sponsored link that appears below an answer: a headline, a destination link, and a brand favicon. That text unit is what the pilot actually runs today; OpenAI has expanded into shopping-style placements, but richer formats like image, video, or interactive conversational units are not confirmed as live, so don't build a campaign expecting them. Write for the conversation - the card should read like the useful next step, not a billboard.
A few practical creative notes: keep the headline concrete and specific to the query, make the favicon recognizable, and use an image only when it genuinely adds information rather than decoration. The card competes with a helpful answer directly above it, so anything vague or salesy gets ignored.
Whatever formats are live when you log in, the rule holds: the ad sits below the organic answer and never alters it. Design for a reader who just got a helpful answer and is deciding what to do next.
Budgets and bidding
You set campaign budgets and choose a bidding model. ChatGPT ads launched around a $60 CPM in February 2026 and fell to roughly $25 within about ten weeks, then added CPC with reported bids of about $3 to $5 per click. The early spend minimum (reported at $200,000, then $50,000) was removed when self-serve opened, so there's no floor to clear. Set a daily cap, start conservative, and confirm current pricing in the interface rather than trusting an old figure.
Setting up your first campaign, step by step
Once your account is verified, the path to a live ad is short. Most advertisers follow the same flow in the manager today:
- Create a campaign and pick an objective, starting with awareness or traffic.
- Set a daily budget and a bidding model: CPM for reach, CPC for clicks.
- Define relevance, the topics and query context where your ad should appear.
- Build the Sponsored link: headline, destination link, and favicon.
- Add tracking parameters to your destination URL so you can attribute results.
- Review the preview, then submit for review. Approval times vary while the platform is in beta.
Keep the first campaign deliberately small and specific. One clear objective, one tight set of queries, and two or three creative variations will teach you more than a broad launch, and it keeps your spend readable while you learn how the auction behaves.
Reporting and measurement
The reporting view shows impressions, clicks, and cost, with more downstream metrics arriving as the platform matures. For now, close the loop yourself: tag your links with UTMs, send clicks to query-specific landing pages, and reconcile ChatGPT-sourced conversions in your own analytics. Don't judge the channel on impressions alone.
A few numbers are worth watching from day one. Click-through rate tells you whether the card earns attention, cost per click tells you whether the auction is affordable for you, and your own conversion rate tells you whether the traffic was worth buying. If the manager and your analytics disagree, trust your analytics and the revenue.
What's rolling out next
The manager is still filling in. In June 2026, OpenAI published Ad Tools Terms that formalize custom audiences and AI-assisted creative tools, and it began a multi-advertiser placement test using second-price auction pricing. These are announced rather than broadly available, so what you can actually use depends on your account and market. The practical takeaway: check what's live when you log in, and don't assume a feature exists just because it's been described publicly.
How it compares to Google and Meta
If you already run Google Ads or Meta, most of the manager will feel familiar, and that's deliberate. The account, campaign, and ad hierarchy is the same, budgets and bids behave the way you expect, and the reporting columns are recognizable. The real differences are two: targeting is built on conversation context and query relevance instead of audience profiles, and the platform is younger, so there are fewer objectives, thinner attribution, and more frequent changes. Treat it as a capable but early cousin of the platforms you know, not a foreign system.
What's still missing
Be realistic about the beta. Objectives are limited, attribution is thinner than on mature platforms, and features change month to month. That's the trade-off for being early: lower competition and cheaper attention in exchange for rougher tooling. If you need a stable, fully instrumented platform today, wait a quarter. If you can tolerate iteration, the learning compounds.
Tips for getting more out of the Ads Manager
- Name campaigns clearly by objective, query theme, and date so reporting stays legible as you scale.
- Change one variable at a time, creative or bid or targeting, so you know what moved the result.
- Refresh creative before fatigue sets in; a card that worked last month can go stale as more brands enter.
- Export data regularly, since beta reporting views can change without notice.
- Check for new objectives and formats each time you log in, and test them small while competition is thin.
Get help running it
We run the ChatGPT Ads Manager day to day for brands, and pair it with organic AI visibility so you show up in both paid and unpaid answers. See our ChatGPT ads management service, or read the step-by-step guide to advertising on ChatGPT if you're setting it up yourself.
