ChatGPT Ads. The Opportunity Is in the Conversation.

The biggest change ChatGPT brings to advertising is that advertising can now appear in the middle of a conversation.

For years, digital advertising relied on a relatively familiar model: identify an audience, select keywords or interests, deliver a message, and measure the clicks. Search engines captured existing demand. Social platforms helped brands create it.

ChatGPT introduces a different context. People use it to explore ideas, compare options, understand problems, and make decisions. That creates a new opportunity for advertisers but it also changes the rules of the game.

When advertising enters a conversation, relevance is no longer just about who sees the message. It is about whether the message makes sense at that moment.

The question is how to become relevant without interrupting the experience that brought someone there in the first place.

A New Place in the Customer Journey

ChatGPT Ads appear below responses, clearly labelled as sponsored content and separated from the assistant’s answer. Depending on the format, an ad can include a brand name, logo, headline, description, image, and link to a landing page.

Ads are shown only to logged-in adult users (18+) on the Free plan and on ChatGPT Go.¹ Plus, Pro, Business, Enterprise and Edu subscribers see no ads, and neither do Temporary Chat sessions. Free users can also switch ads off in exchange for a lower daily message limit.²

The format may look familiar. The context does not.

Someone asking for help choosing running shoes is in a different mindset from someone scrolling through social media. Another may only be trying to understand a problem.

These are different moments, with different expectations. The value of advertising lies in recognising that difference.

For B2B, this means negative selection. Corporate decision-makers mostly work on Enterprise, Business or Plus accounts, where no ads appear, so reaching them through ChatGPT Ads is structurally limited.

ChatGPT Ads are still an evolving product, and availability depends on the market, account type, and current rollout. Advertisers should check OpenAI’s latest documentation before planning a campaign.

Targeting Is Moving Beyond Keywords

Traditional search advertising starts with the words people type. ChatGPT Ads introduce another way to describe relevance: context hints.

Instead of targeting only a phrase such as “running shoes”, a brand can describe a situation: cushioned running shoes for beginners training for their first 5K. The difference is subtle but important. The advertiser is describing a need, not just a search term.

Context hints help the system understand when an ad may be relevant. They do not guarantee placement or work exactly like traditional keyword targeting. They are set at the ad group level, not on individual ads. The technical limit is 2,000 hints per ad group.³ In practice, advertisers use 5 to 15 descriptions of a situation, one sentence each.⁴

Ads Manager does not report clicks or conversions for individual context hints. It aggregates results for the whole ad group.⁴ Editing one hint during a campaign therefore disturbs the test data collected so far, so different intents have to be isolated in separate ad groups.

Ads compete in a relevance-weighted, second-price auction.⁵ The system assesses how well the context hints, the headline, the ad copy and the scanned content of the landing page fit together.⁶

Campaign controls and targeting options also vary, so advertisers should not assume they can reproduce every familiar search or social media setup.

The strategic implication is clear: brands need to understand the problems their customers are trying to solve, not just the keywords they use.

A product catalogue tells you what a brand sells. Customer insight tells you when that product matters.

The New Creative Challenge: Be Useful, Not Just Visible

In a crowded advertising environment, brands have learned to compete for attention. In a conversational environment, attention alone may not be enough.

A generic headline and a vague promise are unlikely to make a strong case for a product. The message needs to communicate a concrete benefit and connect naturally with the need being addressed.

This puts pressure on three things:

• Specificity. Explain what the product helps people do, rather than relying on broad claims.
• Relevance. Build different messages around different customer needs and situations.
• Consistency. Make sure the landing page delivers on the promise made in the ad.

For example, a software brand could develop separate messages for teams trying to save time and businesses looking to simplify reporting. Each message would address a different problem and lead to a relevant landing page.

The objective is not to produce endless variations of the same ad. It is to discover which customer needs create a meaningful connection with the offer.

The Limits Matter as Much as the Opportunity

Every new advertising platform comes with a learning curve. ChatGPT Ads is no exception.

Advertisers should not assume that targeting, remarketing, exclusions, or measurement work exactly as they do on established platforms. After the US launch in February 2026, the platform went live in 31 European markets, including the EU and the wider EEA, at the end of August 2026, with self-serve access at ads.openai.com.⁷ Country, currency and time zone are locked when the account is created and cannot be changed.⁸ Agencies cannot create accounts on behalf of clients.¹

The minimum daily budget is USD 25 (or EUR 15 / GBP 15). The daily budget is settled as a 7-day average, so on a single day the system can spend up to 200% of it.⁹

Billing is per click (CPC) or per thousand impressions (CPM), never directly per conversion, including in conversion-optimised oCPC and oCPM campaigns.¹⁰

OpenAI recommends a starting maximum CPC bid of USD 3 to 5. At the USD 25 minimum, that buys only 5 to 8 clicks a day, so statistically meaningful results need a higher budget.¹¹

Campaign eligibility also depends on OpenAI’s advertising policies, which restrict certain products, services, and types of claims.

Privacy is another important part of the model. OpenAI states that advertisers do not receive users’ private conversations or personal details through the advertising platform. OpenAI’s own algorithm, however, uses earlier chats and ChatGPT’s long-term memory to match ads when a user has Personalized Ads switched on. Personalised ads are not available in the EEA or Switzerland.¹² Ads are also intended to remain separate from the answers generated by ChatGPT.

For marketers, these are major technical details. They shape the campaign strategy, the data available for optimisation, and the expectations that should be set before investing.

The practical lesson is simple: understand the limitations before building a plan around capabilities the platform may not offer.

See OpenAI’s advertising policies for the current requirements.

Clicks Are Not the Whole Story

ChatGPT Ads can provide campaign metrics such as impressions, clicks, spend, click-through rate, and cost per click. Advertisers can also use supported conversion-tracking tools and UTM parameters to connect campaign activity with website outcomes.

Conversion tracking rests on the Measurement Pixel, which captures the oppref URL parameter and sets the __obref cookie, and on the server-side Conversions API (CAPI).¹³ Deduplication requires the same Pixel ID, event name and event_id.¹⁴ A conversion campaign can optimise for one standard conversion event only.¹⁰

In the EEA, the OpenAI Measurement Pixel initialises with consent set to granted by default. Under GDPR and ePrivacy, the script has to be blocked manually (consent: false) until the user agrees in the cookie banner. Events blocked before consent are not replayed afterwards.¹⁵

But the same principle applies here as elsewhere in digital marketing: a click is not a business result.

The more useful questions are whether the campaign generates qualified leads, drives purchases, or brings customers at a sustainable cost. Where revenue data is not available in platform reporting, advertisers need to connect campaign performance with their own analytics and sales data.

Early campaigns should be treated as experiments. Test different messages, context hints, and landing pages. Give each variation enough budget to generate useful evidence, then compare results against other channels.

A new platform does not automatically deserve a place in the media mix. It needs to earn one.

Start With a Need, Not a Format

ChatGPT Ads may be particularly interesting for brands whose customers research options, compare alternatives, or look for solutions before buying. But simply transferring an existing campaign into a new channel is unlikely to be enough.

A better starting point is to identify a specific customer need, develop a clear message around it, and build a landing page that helps the user take the next step.

From there, run a controlled test. Set one clear objective, use a small number of genuinely different creative approaches, and measure the outcomes that matter to the business.

The platform will continue to evolve, so current targeting options, budgets, measurement features, and eligibility rules should be verified before launch.

Relevance Is the Real Advantage.

ChatGPT Ads gives brands a new way to reach people while they are actively exploring questions and decisions. But the presence of a new advertising format does not, by itself, create a new competitive advantage.

That advantage comes from understanding the customer well enough to offer something genuinely relevant, and from measuring whether that relevance translates into results.

In conversational advertising, the opportunity is not simply to enter the conversation. It is to have a good reason to be there.

Less interruption. More relevance.

That is the principle worth carrying into the next phase of digital advertising.