ChatGPT Ads 2026: How Self-Serve Works for Brands

ChatGPT ads 2026 are paid placements shown to people using ChatGPT’s Free and Go plans. Businesses can create campaigns in OpenAI’s self-serve Ads Manager, but buying an ad does not improve a brand’s chance of being cited in an organic ChatGPT answer. Treat ads as a separate paid channel, and build organic visibility through useful, credible content alongside it.

That distinction matters if you are planning a media budget. This guide explains what the platform offers, how to assess whether it suits your audience, and how to measure it without confusing paid exposure with AI search visibility. If you are building an organic programme too, explore SearchFIT’s AI search resources and practical content strategies for AI search visibility.

What are ChatGPT ads in 2026?

ChatGPT ads are sponsored placements that appear in the ChatGPT experience. They are separate from the answer generated by the assistant: the answer is not rewritten to include an advertiser, and a sponsored card is labelled as advertising. OpenAI’s published advertising approach says advertising does not influence the answers ChatGPT gives.

In practical terms, think of the experience as two separate parts. First, ChatGPT produces an answer based on the user’s question and the information available to it. Then, where an ad is eligible and relevant, a clearly marked placement may appear below that answer. The ad is a paid opportunity to reach someone in a relevant context. It is not a citation, recommendation or endorsement from the assistant.

OpenAI announced testing ads in ChatGPT in January 2026, then began live testing in the United States in February. It introduced self-serve Ads Manager in May and expanded access to additional markets in October. OpenAI’s announcement of the October expansion says the platform added 26 countries and that more than 100,000 advertisers had used it. The expansion brought availability to 89 countries, according to the announcement.

Availability is not the same as universal access. A business needs to be in an eligible market and satisfy the platform’s account and advertising requirements. The ads are currently restricted to ChatGPT’s Free and Go tiers. OpenAI’s Ads in ChatGPT help page explains where ads may appear, how users can manage their experience and which subscriptions are ad-free. Check that page before planning a campaign because product access and settings can change.

What does “self-serve” mean?

A self-serve advertising platform lets a business set up and manage campaigns directly through an online interface, rather than buying every placement through a sales representative or agency deal. OpenAI’s platform is called Ads Manager and is available at ads.openai.com. The advertiser controls its campaign details, budget, creative and performance checks in the account.

Self-serve does not mean “set it and forget it”. You still need to decide what you want the campaign to achieve, make sure the offer suits the audience, set up measurement, and review results. As with any new channel, you should also check eligibility and the latest platform rules before committing budget.

Who can see ChatGPT ads?

The audience is shaped by the subscription plan. Under the current arrangement described by OpenAI, ads can appear for users on Free and ChatGPT Go. Paid plans such as Plus, Pro, Business, Enterprise and Edu are ad-free. That makes the placement different from a channel that reaches every person using the product.

This distinction is important for business-to-business marketers. A potential buyer using a company-managed Business or Enterprise workspace will not see these ads. A person researching software on a Free or Go account might. That means ChatGPT ads may be more relevant for reaching consumers, small businesses, individual professionals and people exploring products on a free account than for reaching an entire enterprise buying committee.

There is a further audience limit: users under 18 do not receive ads, and ads are not placed in certain sensitive or vulnerable conversations. OpenAI’s ad policies set out prohibited and restricted categories, along with requirements for advertising in regulated areas. Users on Free may also have the option to turn off ads in exchange for tighter daily usage limits, as described in OpenAI’s help documentation.

So, do not estimate reach from ChatGPT’s total user base alone. The usable audience is the portion of users who are in an eligible market, on an ad-supported tier, in an eligible context and able to receive the type of ad you plan to run. The Ads Manager’s campaign estimates and actual delivery will be more useful than a broad platform-wide audience figure.

How ChatGPT Ads Manager works

Ads Manager is the campaign workspace for eligible advertisers. OpenAI describes new buying options, including self-serve, in its announcement about ways to buy ChatGPT ads. At a high level, a marketer creates an account, chooses a campaign objective, sets its audience context and budget, supplies ad creative, then monitors delivery and results.

The interface and available options may evolve, so treat any checklist as a planning framework, not a promise that every setting is available to every account. Before launch, confirm eligibility, billing requirements, bid options and tracking support in Ads Manager itself.

1. Set up an eligible business account

Start by checking whether your business and target market are supported. The account may require business details and payment verification. You will need a payment method for billing, and your ads must comply with OpenAI’s content and category rules.

If you work at an agency, settle account ownership and access before building campaigns. Decide whether the client or agency will own billing, who will be able to edit the ads, and who is responsible for policy checks and reporting. Keep each client’s permissions and conversion data separated appropriately.

2. Choose an objective and budget

The campaign objective should match the action you want a person to take. That might be visiting a product page, signing up for a trial, buying a product or completing another measurable conversion. A conversion is a valuable action taken after an ad interaction, such as a purchase or form submission.

The minimum daily budget reported for a self-serve campaign is US$25 per campaign, as described in Everything-PR’s ChatGPT ads guide. Treat that as a minimum, not as evidence that the budget will produce meaningful results. You need enough spend and time to gather data, and the required amount depends on your goal, audience, conversion rate and product economics.

OpenAI has described buying options that include cost per thousand impressions (CPM), cost per click (CPC) and outcome or conversion optimisation. CPM means paying based on ad impressions, or how often the ad is shown. CPC means paying for clicks. With outcome optimisation, the system aims to find delivery more likely to produce the selected action. Check which options are available for your campaign and market in the account before planning around them.

3. Define the context you want to reach

ChatGPT ads do not use the familiar search-ad system of bidding on exact, phrase or broad-match keywords. Instead, advertisers provide contextual hints that describe what the product is for and which kinds of conversations could be relevant. The system considers the current prompt and conversational context to assess whether an ad fits.

For example, a meal-kit company might describe its offer as relevant to people looking for quick weeknight meal ideas, rather than trying to buy a particular phrase. A project management SaaS business might focus on teams comparing ways to organise work. These descriptions are not a guarantee that an ad will appear for every related question. They help explain relevance; the platform controls matching and delivery.

This is one reason to write your campaign brief in terms of the user’s task. What are they trying to solve? What stage are they at? What makes your offer useful in that moment? A clear answer helps you shape the ad and landing page, even if the platform’s matching controls differ from a keyword account.

4. Prepare creative and the destination

An ad is only useful if the message and landing page agree. Make the offer clear, state who it suits, and send people to a page that delivers on the promise. If the ad talks about a specific product, the destination should make that product easy to understand and buy. If it promises a demo, the page should make booking one straightforward.

Use claims that you can substantiate. Keep the title and description specific enough for someone to understand the value without treating the assistant’s answer as part of your creative. Make sure the page works well on mobile, explains key terms and gives the visitor a clear next step.

5. Add conversion measurement

OpenAI has described support for an advertising pixel, a Conversions API (CAPI) and partner integrations. A pixel is a small piece of tracking code on a website that can report selected events, such as a purchase. CAPI sends event data from a business’s server to the ad platform. Partner integrations can help connect campaign and website data without building every connection from scratch.

Set this up before launch if conversions are the goal. Decide which events matter, check that the events fire correctly and follow applicable privacy and consent rules. Avoid collecting or passing data you do not need. OpenAI’s help documentation explains its ad experience and information handling; consult it alongside your own privacy obligations before configuring measurement.

6. Review delivery and results

Ads Manager reporting can include impressions, clicks, CPC and conversion events, alongside diagnostic or event-quality information. An impression tells you an ad was displayed. A click indicates someone selected it. A conversion records a selected action after an interaction, subject to the tracking setup and attribution rules.

Do not treat a high click-through rate as proof that a campaign is profitable. Look at the full path: whether the right people clicked, whether the landing page matched their needs, whether they completed the intended action, and whether the value of that action justifies the spend. Compare results with a relevant baseline from your own business, not an assumed industry average.

ChatGPT ad formats and where they appear

The format determines how your message is presented. Published reports suggest current tests and options include sponsored cards, product catalogues and carousels, visual ads around image generation, and sponsored agent experiences. Not every format will be available in every market or account. OpenAI’s help page is the best place to confirm what is currently live.

Sponsored cards

A standard sponsored card appears in a clearly separated area below the generated response and carries a “Sponsored” label. The card can include a brand name, a title, a short description, a link and an image. It is not inserted into the assistant’s answer as though the model had independently recommended the advertiser.

This format is suited to a focused message and a clear next step. For instance, a retailer could promote a relevant product category after a shopping-related conversation. A software company might invite someone to compare plans or start a trial. The card does not replace the need for a useful destination page or a well-defined offer.

OpenAI has also described automated creative adjustments, such as testing headline variations, adapting images and translating copy for a user’s locale. Where those tools are offered, review how the final versions represent your brand. Automation can help with variations, but the advertiser remains responsible for accurate claims, suitable language and a landing page that matches the message.

Product catalogues and carousels

A product catalogue is a structured feed of the items a business sells. It can include information such as product names, images and destination links. A carousel is a set of cards that a user can browse, rather than a single static placement. This format is aimed at product discovery and shopping-related conversations.

E-commerce teams should pay particular attention to feed accuracy. Keep availability, pricing and product details aligned with the store. If the ad shows an item that is out of stock or sends people to an unrelated page, the experience becomes frustrating and measurement becomes harder to interpret. Brands using Shopify can review SearchFIT’s e-commerce resources when planning how product information and organic discovery support the wider customer journey.

Visual ads in image-generation experiences

OpenAI has announced tests of visual placements in image-generation workflows. In the described format, a user creating images may see a set of square lifestyle or product images. This is a distinct context from the sponsored card below a text answer. It may suit visual products, but an advertiser should first confirm availability and understand how the placement is labelled and shown.

The key planning question is not simply whether your product looks good in an image. Ask whether the user’s task creates a natural reason to see it. A relevant visual can help someone explore a style or product. A placement that interrupts an unrelated creative task risks feeling intrusive and may not attract useful attention.

Sponsored agents

A sponsored agent is a “click-to-chat” experience that takes a user into a conversation with a brand’s verified AI agent. Rather than sending someone straight to a conventional landing page, the ad invites them to ask questions or explore an offer in a chat interface. These experiences have been described as pilots, so do not assume the feature is broadly available.

If your business considers this approach, plan the agent as part of the customer experience, not just as an ad format. It needs accurate product information, clear boundaries, a useful way to handle uncertainty and a sensible hand-off to a person where needed. A conversational route can remove friction for some enquiries, but it also creates responsibility for what the agent tells people.

How targeting differs from keyword advertising

In a search ad platform, an advertiser may choose a keyword and specify how closely a user’s search must match it. ChatGPT ads use a different model. The advertiser describes the product, category and relevant situations; the platform assesses the conversation’s context and intent.

That difference changes the planning work. You cannot rely on a list of competitor terms or a set of exact-match queries to define your campaign. Nor should you assume the system provides the same audience interest categories or lookalike controls that are familiar from social advertising. Build a clear description of the user need, then test whether the campaign reaches suitable contexts.

A practical contextual brief can answer three questions:

• What is being offered? Describe the product or service in plain language, including what it does and what makes it distinct.
• Who is it for? Name the kind of person or organisation that benefits, without trying to infer sensitive traits.
• When is it relevant? Explain the task, problem or decision that could make the offer useful.

Consider a home furnishings retailer. Instead of trying to buy a phrase such as “best sofa”, describe the products and the kinds of living-room planning or furniture-shopping conversations they fit. For a SaaS brand, describe the workflow problem and the type of team the software serves. Then make the ad and destination page answer the same need.

Personalisation can also be part of the experience. OpenAI says that, for logged-in users who have ad personalisation enabled, broad conversational patterns and previous ad interactions may inform relevance. Advertisers do not receive access to individual conversations, prompt histories or stored memories. They receive campaign reporting, not a transcript of what a particular person asked.

That privacy boundary matters to campaign design. Do not build a strategy around seeing someone’s prompts or targeting a person based on a private conversation. Use the campaign settings and reporting available in the account, and design the message around a legitimate product need rather than an assumption about an individual.

What ChatGPT ads cannot do for organic visibility

Paid advertising and organic AI visibility are separate systems. Organic visibility means your brand, content or website is found and used in an answer without buying an ad placement. In a ChatGPT search experience, that may include a source citation or a link to a web page. Paid visibility is a sponsored placement delivered through an advertising campaign.

The separation is central to OpenAI’s advertising principles. Paying for an ad does not buy a place in ChatGPT’s answer, improve your organic citation rate, change the assistant’s ranking of sources or give your website a special position in the model’s responses. Likewise, an organic citation does not mean you have paid for an ad. The two outcomes should be measured independently.

A simple example shows the difference. Imagine a person asking ChatGPT to compare accounting tools for a small business. The assistant may produce an answer and cite relevant sources it has found. A software brand could also qualify for a sponsored card beneath the answer if its campaign is eligible and relevant. The citation is an organic outcome; the card is a paid one. One does not cause the other.

This means a brand can use both channels, but should give each a distinct job. Ads can support immediate, measurable promotion, such as a product launch or a time-limited offer. Organic work helps make your expertise and product information available when AI systems and people look for answers. A combined approach may give a brand two separate opportunities to be noticed, but there is no guarantee that both will appear in the same response.

For organic discovery, focus on making important pages easy to find, accurate and clear. Explain what your product does, who it is for, what it costs where appropriate, and how it compares with alternatives. Publish useful answers to real customer questions. Earn credible mentions and citations from relevant third-party sources. Use structured data where it accurately describes the page, and keep the information on your site current.

Those are not shortcuts to a citation. AI systems decide what to use, and no publisher can guarantee a mention. But they are durable content and site practices that can support visibility across search experiences. SearchFIT’s guide to AI search content strategies covers practical ways to build that foundation. You can also use SearchFIT’s comparison page to understand how a dedicated AI visibility workflow fits alongside other marketing activity.

How to decide whether ChatGPT ads fit your brand

Start with the audience rather than the novelty of the platform. Ask whether your likely buyer uses an ad-supported ChatGPT plan, whether the product fits the kinds of conversations where ads are eligible, and whether you have a strong next step for someone who is still researching.

For an e-commerce brand, the channel may be worth testing if products are relevant to active shopping or product-discovery conversations and the catalogue is accurate. For a consumer subscription, the ad could promote a trial or explain a useful feature. For a B2B SaaS company, consider whether individual professionals or small teams are a meaningful segment of the addressable audience. If your strategy depends on reaching buyers inside company-managed Enterprise workspaces, these ads will not reach those users under the current policy.

Then assess the economics. Decide what a click or conversion is worth to your business. Estimate the margin or long-term value of a new customer using your own data, not an unverified platform benchmark. Set a test budget that your team can afford to learn from, and agree in advance what evidence would justify continuing, changing or stopping the campaign.

A useful test plan includes:

• One clearly defined audience context and offer.
• A landing page designed for that offer, rather than a generic homepage.
• Correct conversion tracking and a privacy review before launch.
• A budget and test period that your finance and marketing teams have approved.
• A review of delivery, clicks, conversion quality and downstream value.
• A note of what changed during the test, so you can interpret the result honestly.

Do not judge a new channel by impressions alone. Impressions show delivery, not business impact. Clicks show interest in the ad, not necessarily purchase intent. Conversions are more informative, but only if the event is configured correctly and the action has real value. If someone clicks but does not convert, investigate the offer, the destination page, product fit and tracking before assuming the platform is at fault.

Brand safety, privacy and policy checks

ChatGPT ads are not available in every kind of conversation. OpenAI’s ad policies restrict placements in sensitive or emotionally vulnerable contexts, including health, mental health, addiction and bereavement. The policies also prohibit some categories, including political ads, weapons, illegal goods and unregulated financial products. Certain regulated advertising, including health, legal and finance, has extra restrictions and verification requirements.

Read the current policy before developing creative or choosing a campaign category. A product that is legal to sell is not automatically eligible to advertise in every market or context. Keep claims factual, avoid implying that the assistant endorses your product, and make sure any regulated claims and disclosures meet the applicable rules.

Privacy deserves equal care. The platform’s contextual approach does not give advertisers a window into a person’s conversation. Use aggregated campaign information to assess performance and follow your own data-protection obligations when sending conversion events. Make sure your privacy notices and consent mechanisms reflect the tracking you use, and avoid collecting personal information that is not necessary for the stated measurement purpose.

If you work with an agency, write down who checks policy, who approves the final ads and who responds if a campaign is rejected or paused. A documented process helps avoid preventable delays and gives the client a clear record of decisions.

A sensible way to combine paid and organic work

Think of ads as a way to meet a relevant person at a particular moment. Think of organic AI visibility as the result of making your business and its information discoverable and useful across many questions. Both can be valuable, but they have different inputs, costs, timelines and measures of success.

A practical sequence is to establish the basics of organic visibility, then run a paid test where the audience and offer make sense. Keep the measurement separate. Use campaign reports to understand paid delivery and conversions. Track organic mentions and citations separately across relevant AI search experiences. Do not report an ad view as an organic recommendation, or a citation as paid reach.

If both channels perform, you can plan them together without claiming one caused the other. For example, an organic source citation could build confidence while a separate sponsored card promotes a trial. That is a useful “surround sound” scenario, but it is not guaranteed and should not be the only reason to buy ads. Your budget decision should still stand on its own expected paid performance.

For a wider marketing plan, consider how the ad’s landing page, product information and organic content work together. A helpful article may answer a research question; a product page may show the exact offer; an ad may introduce that offer in a relevant context. Consistency across those touchpoints makes it easier for a person to understand what you do, whether they arrive from an ad or an AI-generated answer.

The bottom line on ChatGPT ads in 2026

ChatGPT ads are a developing paid channel with self-serve campaign management, contextual matching and a range of formats. They are available to eligible advertisers in supported markets and appear to users on Free and Go plans, not paid ChatGPT plans. Availability, formats and settings can change, so confirm the current details with OpenAI before launch.

Most importantly, advertising is not a shortcut to organic AI search visibility. It can buy sponsored exposure when the context and audience fit. It cannot buy a citation or influence the answer itself. Test ads on paid-channel economics, and build organic visibility with useful content, clear product information and credible sources. To plan the organic side, start with SearchFIT’s resources for AI search visibility and choose a measurement approach that keeps paid and organic results distinct.