OpenAI is expanding ChatGPT advertising with new visual formats, additional brand safety controls and improved measurement capabilities, as it continues developing its advertising business.

The new placements will allow advertisers to showcase products, services and experiences through imagery, initially appearing alongside images generated by ChatGPT users.

Testing will begin later in October with a selected group of US advertisers. OpenAI says advertisements will be clearly labelled, remain separate from generated images and have no influence over ChatGPT’s responses.

The announcement follows the introduction of ChatGPT advertising in February and a series of subsequent developments covering campaign management, creative tools and commercial integrations.

With 1.2 billion weekly users, ChatGPT provides OpenAI with a substantial potential advertising audience. However, advertisements are currently limited to eligible users on its Free and Go plans, meaning the addressable audience is considerably smaller than its total user base.

OpenAI is entering a highly competitive market dominated by Google, Meta and other established advertising platforms. While early advertisers have reported some encouraging results, questions remain around available inventory, campaign measurement and how effectively advertising within AI conversations can influence commercial outcomes.

ChatGPT Visual Ads Expand Beyond Sponsored Placements

ChatGPT’s existing advertising products include sponsored placements displayed beneath responses, combining familiar advertising elements such as headlines, descriptions, images and links.

The latest format introduces advertising into ChatGPT’s image generation experience, allowing advertisers to use imagery to demonstrate products, illustrate their use and showcase relevant experiences.

A furniture retailer, for example, could potentially promote a product to someone generating interior design ideas. A travel business might showcase destinations or experiences related to the images someone is creating.

Those are potential applications rather than confirmed targeting capabilities. OpenAI has yet to provide detailed information about the new format’s targeting controls, pricing or expected advertising inventory.

The initial rollout will also be limited to the United States, with additional formats expected as OpenAI develops its advertising offering.

Visual advertising extends the commercial possibilities beyond conventional search-style placements. ChatGPT users frequently explore ideas, compare alternatives and refine their requirements through conversations, potentially providing advertisers with opportunities to reach consumers during earlier stages of consideration.

However, the effectiveness of advertising within these activities remains largely unproven at scale.

OpenAI Pushes Further Into Conversational Commerce

OpenAI’s advertising ambitions extend beyond introducing additional placements.

The company has been expanding its advertising capabilities since February, adding self-service campaign management, additional buying options and more sophisticated creative formats.

Its advertising system uses conversational context to identify potentially relevant placements, introducing some interesting differences from conventional search advertising.

A Google search for running shoes typically provides a clear commercial signal. Someone discussing a marathon training programme with ChatGPT may also have an interest in running shoes, although purchasing equipment could be only one aspect of a much wider conversation.

The distinction creates both opportunities and complications for advertisers.

ChatGPT users can explore unfamiliar products, compare competing brands, investigate specifications and refine their preferences through successive interactions.

Advertising appearing within those activities could potentially influence consideration before a consumer begins actively searching for somewhere to purchase.

OpenAI has also been exploring experiences that allow users to move from sponsored placements into conversations with businesses, extending the potential relationship between advertising, product discovery and customer engagement.

However, advertisers will need to distinguish between general conversational relevance and genuine commercial intent. A product mentioned during an AI conversation does not necessarily indicate that someone is ready to buy.

The relationship between advertising and organic recommendations also requires careful consideration.

OpenAI maintains that paid placements remain separate from ChatGPT’s generated answers. Advertisers cannot purchase favourable recommendations within the assistant’s responses, even when their advertisements appear alongside relevant conversations.

ChatGPT Advertising Faces Questions Around Inventory And Performance

OpenAI’s advertising business has attracted considerable attention, although early advertiser experiences have been mixed.

According to Digiday, advertisers have raised three recurring concerns since ChatGPT Ads launched: limited measurement capabilities, relatively few advertising formats and insufficient inventory to accommodate larger budgets.

The introduction of visual placements addresses the range of available creative formats, while OpenAI’s expanding measurement partnerships respond to concerns about campaign reporting.

Advertising inventory presents a different challenge.

ChatGPT may have enormous overall reach, but advertisements are restricted to eligible users and suitable conversational contexts.

The number of available placements also depends on how frequently users encounter situations where OpenAI considers advertising appropriate.

Large advertisers accustomed to substantial spending across Google and Meta therefore cannot assume they will find comparable opportunities to deploy budgets within ChatGPT.

Early performance evidence has also been inconsistent.

An independent collection of 27 publicly reported ChatGPT advertising results, compiled by MadContext, found reported costs per click ranging from US$1.72 to US$13 across the 14 campaigns where that information was available.

The same collection found that all five campaigns comparing ChatGPT’s reported clicks with their own analytics recorded fewer website visits than platform-reported clicks.

Differences between recorded advertising clicks and website sessions can arise for several reasons, including visitors abandoning pages before analytics loads, tracking prevention, consent settings and different counting rules.

However, the findings illustrate some of the practical difficulties advertisers face when attempting to evaluate an emerging channel.

The MadContext analysis also aggregates publicly disclosed results rather than a representative sample of ChatGPT advertisers. Its relatively small sample and varied campaign conditions limit the conclusions that can be drawn about typical performance.

ChatGPT Advertising Studies Show Mixed Results

Published conversion data presents an equally mixed picture, although the scope and quality of the available evidence vary considerably.

Some early studies draw on relatively small samples, while others focus on specific industries, campaign types or advertisers. Several combine observed campaign results with estimates based on established digital advertising benchmarks.

OpenAI’s own published case studies have highlighted favourable outcomes from selected advertisers, although individual campaigns provide limited evidence of how the platform performs across the wider market.

A successful campaign for a specialist ecommerce brand, for example, offers relatively little guidance for a university, financial services organisation or B2B advertiser operating with longer purchasing cycles.

Differences in measurement methods complicate comparisons further.

Some studies report click-through rates or attributed conversions, while others examine incremental sales or customer acquisition costs.

An advertisement generating relatively expensive clicks could still deliver profitable customers. Equally, a high conversion rate offers limited insight into commercial performance without considering acquisition costs and customer value.

Small sample sizes, narrow industry coverage and inconsistent measurement methods make it difficult to establish reliable ChatGPT advertising benchmarks.

Larger independent studies covering a broader range of advertisers, campaign objectives and industries will provide a clearer assessment of the platform’s commercial effectiveness.

However, conventional advertising performance measures introduce another complication. ChatGPT’s influence on consumer decisions may extend considerably beyond the clicks and conversions typically captured in campaign reporting.

AI Recommendations Are Already Influencing Brand Consideration

ChatGPT and other AI assistants are increasingly being used to research products, compare competing businesses and evaluate purchasing options.

Their responses can influence which brands consumers consider, potentially shaping decisions long before someone clicks an advertisement or visits a website.

Research published in June 2026, From Prompt to Purchase, examined the relationship between AI-generated brand recommendations and subsequent consumer behaviour.

The researchers connected opt-in browsing data with conversations across ChatGPT, Claude and Gemini, examining what happened after users encountered brand recommendations.

Among users with no recent observed engagement with a recommended brand, the study identified:

• A 4.3 percentage point increase in subsequent branded Google searches.
• A 2.4 percentage point increase in visits to the brand’s website.
• A 1.0 percentage point increase in visits to brand-specific retailer pages.

The observational research does not establish definitive causality or completed purchases. It was also conducted by researchers associated with Scrunch AI, a company operating in the AI visibility measurement market.

Nevertheless, the findings provide evidence of a relationship between AI recommendations and subsequent consumer behaviour outside the original conversation.

A consumer might encounter a brand through ChatGPT, develop an interest in its products and subsequently search for it on Google before purchasing.

Conventional attribution could credit the resulting conversion to Google Search, direct traffic or another channel, overlooking the earlier influence of the AI recommendation.

The commercial influence of AI assistants may therefore be considerably broader than the referral traffic they generate.

The study examined organic AI recommendations rather than paid advertising, so its findings cannot be used to establish the effectiveness of ChatGPT Ads.

However, the results raise an important question about the types of outcomes advertisers should measure as commercial placements become more closely associated with conversational product discovery.

ChatGPT Advertising Measurement May Miss Wider Brand Effects

OpenAI is responding to advertiser concerns by expanding its measurement infrastructure and introducing integrations with established marketing technology providers.

New partnerships with Hightouch, Tealium and LiveRamp will help advertisers connect existing customer and conversion information with ChatGPT campaigns.

Additional attribution integrations include AppsFlyer, Adjust, Triple Whale and other measurement providers, extending reporting across websites, mobile applications and ecommerce.

OpenAI is also working with Haus, Measured and WorkMagic on incrementality studies designed to estimate the additional business generated by advertising.

The developments address some of the practical limitations advertisers have encountered.

Digiday recently reported that agency Accuracast ran a UK lead generation campaign that generated form submissions, while ChatGPT’s advertising interface continued showing zero conversions two weeks later.

Better integration with advertisers’ existing measurement systems should help resolve some of these reporting discrepancies.

However, much of the emerging infrastructure remains grounded in established advertising metrics, including clicks, conversions and attributed sales.

Conversational AI potentially influences purchasing behaviour much earlier, shaping brand awareness, consideration and preferences before consumers demonstrate a clear intention to purchase.

An advertisement encountered during a broader product research conversation might contribute to a favourable brand impression without generating an immediate click.

Similarly, consumers may subsequently search for a brand, visit its website independently or purchase through a retailer, leaving little direct evidence of the original advertising exposure.

Brand lift studies, changes in branded search behaviour and controlled experiments examining subsequent purchasing activity could therefore provide additional perspectives on campaign effectiveness.

OpenAI’s incrementality partnerships may help address some of these challenges, particularly where experiments can isolate the impact of advertising exposure on later customer behaviour.

However, the influence of paid advertisements must also be distinguished from ChatGPT’s organic recommendations.

Advertisers need to establish whether sponsored placements generate additional demand, reinforce existing preferences or reach consumers who were already likely to purchase.

The relationship between organic AI visibility and paid advertising is still developing, creating a measurement challenge that extends beyond the capabilities of conventional click-based attribution.

OpenAI Expands ChatGPT Brand Safety Controls

OpenAI is also developing additional safeguards around advertising placement.

The company has introduced Negative Phrases for qualifying advertisers, allowing businesses to apply further restrictions to the contexts in which their advertisements may appear.

It is also working with DoubleVerify and Integral Ad Science to develop brand suitability evaluation programmes.

Conversational advertising introduces some distinct brand safety challenges.

Unlike a conventional website or video, the subject of a ChatGPT conversation can change repeatedly as users ask follow-up questions.

A discussion beginning with an everyday consumer question could move into financial difficulties, personal circumstances or other sensitive subjects.

OpenAI says its safeguards assess whether conversations are appropriate for advertising and restrict placements in unsuitable environments.

The independent evaluation partnerships may provide advertisers with additional assurance, although the programmes remain under development.

OpenAI also maintains that sponsored placements remain separate from AI-generated answers and that advertisers cannot access users’ private conversations.

Protecting that separation is particularly important as ChatGPT expands its commercial capabilities. Consumers may place considerable trust in information provided by AI assistants, making transparency around paid placements an important consideration for both OpenAI and advertisers.

ChatGPT Ads Create A Broader Measurement Challenge

ChatGPT advertising is developing quickly, with OpenAI introducing additional creative formats, attribution integrations and brand suitability controls.

The latest developments address several concerns raised by advertisers during the platform’s early commercial rollout.

However, OpenAI’s substantial consumer audience should be considered separately from its available advertising inventory, while published performance results remain difficult to compare.

The growing influence of AI assistants on product research and brand consideration also complicates conventional advertising measurement.

Clicks and conversions will remain important indicators, particularly for advertisers focused on direct response. However, they may capture only part of the commercial influence generated when consumers encounter brands during AI-assisted research.

Advertisers evaluating ChatGPT campaigns will need to consider both immediate performance and the potential contribution to broader brand consideration and purchasing behaviour.

Controlled experiments, incrementality studies and brand measurement may help establish that relationship, although independent evidence remains limited.

For now, ChatGPT represents an emerging advertising environment with considerable potential audience scale, developing measurement capabilities and unresolved questions about commercial effectiveness.

Understanding how paid placements influence consumer decisions, separately from the recommendations ChatGPT generates organically, will be central to evaluating its advertising value.