How to advertise on ChatGPT: eligibility, setup, and campaign launch

To get a ChatGPT ads campaign live, you move in a fixed sequence: confirm eligibility, get an account approved, build the campaign shell, configure ad groups, add ads that can pass review, then interpret early delivery with reporting lag in mind. For high-level eligibility and formats, see the ChatGPT ads overview, and use this guide for the step-by-step build.

Check country and category prerequisites before you start

Before you touch Ads Manager, confirm that the billed legal entity for the account is based in one of 63 supported countries as of October 2026. OpenAI’s documentation states that Ads Manager availability is tied to the country of the billed entity, not the user’s physical location or login IP.

Next, confirm that your business category is eligible. Per OpenAI’s ad policies, launch categories are primarily consumer verticals, while financial, healthcare, and legal advertisers are handled case by case in the US and can require proof of licensure. Several categories are disallowed entirely, so if you are in a sensitive or regulated space, you should validate against policy before planning creative or budgets; for deeper category rules, see the category eligibility detail.

If you are in a regulated financial, healthcare, or legal category outside the US, plan for the possibility that you cannot run or that you will face additional scrutiny. Treat this eligibility check as a hard gate, because no campaign setup work can override a category restriction in review.

Create and verify your OpenAI Ads account

Account creation for ChatGPT advertising starts at ads.openai.com, where you sign up for Ads Manager. According to OpenAI’s policies, new advertisers go through an advertiser review that checks business legitimacy using verification data, account quality signals, and risk indicators.

Make sure the business details you enter match your billing and any public records you control, because inconsistencies can slow or block approval. OpenAI’s documentation notes that some new self-serve accounts can initially advertise only in their home country, with other countries unlocking only after identity verification is approved and a required spend threshold is met from home-country campaigns.

That home-country-only phase means you should plan to launch initial campaigns in your entity’s country even if your long-term plan is multi-country. Do not promise stakeholders immediate global reach until you see in your account that additional geos are available in the campaign builder.

Build the campaign shell with locked and flexible settings

Once your account is active, create a new campaign in Ads Manager. Per OpenAI’s help documentation, a campaign contains a title, objective, optional conversion event (for the Conversions objective), budget, start and end dates, targeted countries, optional platform targeting, and optional custom audiences.

The objective is the first irreversible decision. You choose between Views (billed on CPM), Clicks (billed on valid-click CPC), and Conversions (billed on oCPC per valid click or oCPM per 1,000 impressions), and OpenAI states that you never pay per conversion. OpenAI also states that the objective cannot be changed after creation, so if you pick the wrong objective you cannot “fix it later” inside that campaign; the correction is to create a new campaign with the desired objective.

Next, choose your budget type. OpenAI supports campaign-total budgets, which give strict total control, and daily budgets. Documentation specifies that a daily budget is interpreted as a 7-day delivery target, where a single day can spend up to 2x the daily amount, and the Sunday–Saturday week will not exceed 7x the daily value. This means the daily budget does not cap what a single day can spend; it caps the 7-day total, with day-level flexibility up to 2x.

Some budget-type changes are one-way. Per OpenAI’s guidance, a campaign set to campaign-total can be switched to daily, but once you do that, you cannot switch that campaign back to campaign-total. This contradicts the idea that you can switch budget types freely while testing; you must treat the move from campaign-total to daily as irreversible for that campaign.

OpenAI’s help center specifies a minimum daily budget of $25 USD, with documented currency floors for 24 currencies such as 15 EUR, 15 GBP, and 2,500 JPY. There is no published rule in the research about any other minimums, so when you plan tests, your daily line items must meet or exceed these floors. OpenAI also recommends starting with a daily budget so you can monitor delivery before committing to a campaign-total budget.

Set your campaign dates so that the start date is today or later and the end date matches your flight plan. A campaign that exhausts its weekly allowance early under a daily budget can stop delivering while still marked active, so if you see delivery pause mid-week, check whether the effective 7x weekly cap has been reached.

For geography, OpenAI’s documentation states that geo targeting supports states or regions, cities, markets, and postal codes where available. The exception is product-feed campaigns, which can only target at the country level. If you need sub-country precision, confirm that you are not using a product-feed campaign type before you expect ZIP or city controls to work.

Platform targeting is optional at the campaign level. You can choose among Android app, Android web, Desktop web, iOS app, and iOS web, and later reporting can segment performance by platform. If you have no reason to constrain platforms, you can leave this open and use Insights to identify where performance is strongest.

Custom audiences are also optional. Per OpenAI’s documentation, inclusion audiences must have at least 25,000 matched users before they can be used, while exclusion audiences can be smaller. OpenAI also recommends that campaigns targeting users in the EEA or Switzerland avoid custom audiences while ads personalization is unavailable there, so do not plan EEA audience strategies around these features until that limitation changes.

Configure ad groups with bids and context hints

Inside each campaign, you create one or more ad groups. OpenAI recommends using multiple ad groups per campaign to test different themes, value propositions, or messaging approaches, rather than trying to cover everything in a single group.

Bids are set at the ad-group level as either a max CPM or a max CPC, depending on your campaign objective and buying strategy. OpenAI’s basics guide recommends a starting max CPC of $3–5, and Ads Manager provides bid-strength guidance that indicates whether your bid is likely to be competitive in the auction.

Context hints are also configured at the ad-group level. Per OpenAI’s documentation, these hints guide matching but are not exact-match keywords and do not guarantee delivery against specific queries or prompts. Treat context hints as directional signals to the system about where your ads are relevant, not as hard targeting constraints.

Create ads that can pass policy and landing page review

Each ad in an ad group must include an advertiser name, a favicon or logo, a title, body copy, a landing page link, and at least one image asset. These fields are required for the ad to be eligible to serve, so you should prepare them before you start bulk-creating ads in the UI.

OpenAI’s ad policies specify that every ad’s title, copy, media, and landing page are reviewed before serving. Systems must be able to review the landing page content itself; if OpenAI’s systems cannot review an ad or its landing page, that ad is not eligible to run.

Because review covers the landing page, not just the ad creative, you must ensure that the destination URL resolves correctly, does not block OpenAI’s systems, and is consistent with the claims made in the ad. If you are in a sensitive category such as financial, healthcare, or legal, align both ad and landing content with the stricter rules that apply to those categories as described in OpenAI’s policies.

Launch, troubleshoot delivery, and read early performance

When your campaigns, ad groups, and ads are configured, set their statuses to active and confirm that the campaign start date includes today. If a campaign is not serving, OpenAI’s help documentation recommends checking that account verification and billing are complete, the campaign and ads are active, the campaign dates include the current day, and all ads have passed review.

Once serving begins, OpenAI states that impressions, clicks, and click-through rate refresh roughly every 15 minutes. Spend reporting is slower, with documented lags of 7–8 hours behind impressions and clicks, so a zero-spend readout on day one does not mean that no charges are accruing; it can simply reflect reporting delay.

This reporting lag is where many new buyers misread performance. Do not assume that zero spend in the dashboard means nothing is being billed, and do not pause or radically change campaigns based on the first few hours of impression and click data alone. Wait for spend to catch up and for at least a full day of data before making structural changes like objective or budget-type shifts, which can require new campaigns or are irreversible.

In the first days, focus on confirming that delivery is happening in the intended countries and platforms, that click and impression volumes are consistent with your bids and budgets, and that there are no policy or review rejections. For deeper pricing behavior and test-budget planning, see the ChatGPT ads cost and budgeting guide, and for conversion tracking implementation before you optimize to the Conversions objective, use the measurement setup reference.