ChatGPT Ads: A New Performance Channel Now Available Through Point2Web

Advertising inside AI-powered conversations is quickly developing into a new user acquisition opportunity for brands. As users increasingly turn to ChatGPT to research products, compare services, plan purchases, and solve specific problems, advertisers can now appear during these moments through OpenAI Ads Manager.

However, ChatGPT Ads works differently from traditional search and social advertising. Campaign delivery is driven primarily by conversational context rather than keywords or demographic targeting, while the available formats, eligible verticals, account requirements, and measurement setup follow their own rules.

Following the latest OpenAI updates, the platform now offers broader market availability, product feed advertising, expanded conversion optimization, additional audience capabilities, and deeper tracking integrations. We provide clients with the infrastructure, campaign expertise, and full-cycle support required to assess, launch, measure, and optimize ChatGPT Ads as a new performance channel.

Where Can You Target OpenAI Ads?

OpenAI Ads gives advertisers access to a broad and growing geographic footprint, creating opportunities to test the channel across established Tier-1 markets as well as Europe, the Middle East, and North Africa.

Available GEOs cover 56 markets worldwide, including:

North America & LATAM: United States, Canada, Mexico, Brazil
Europe: United Kingdom, all 27 EU member states, Iceland, Liechtenstein, Norway, Switzerland, and Türkiye
Asia-Pacific: Australia, New Zealand, Japan, South Korea, India
Middle East & North Africa: Algeria, Bahrain, Egypt, Iraq, Israel, Jordan, Kuwait, Lebanon, Morocco, Oman, Qatar, Saudi Arabia, Tunisia, and the United Arab Emirates

Beyond country-level targeting, advertisers can get more granular depending on the market. Campaigns can include or exclude specific locations and target users by state or region, city, designated market, and postal code, giving brands more control over where their budgets are concentrated.

For brands looking to test OpenAI Ads without adding another platform setup to their workload, Point2Web can manage the launch and help shape the GEO strategy around the campaign’s audience, goals, and performance targets.

Who Can See ChatGPT Ads?

Ads are currently shown to eligible users of ChatGPT Free and Go. They are not served to Plus, Pro, Business, or Edu subscribers, users identified as being under 18, Temporary Chats, or Atlas environments.

OpenAI can also show certain all-audience campaigns to logged-out users. In the European Economic Area and Switzerland, additional privacy limitations apply: ads use the context of the current conversation without personalization, and Custom Audiences should not be used for campaigns targeting these markets.

The ads appear below ChatGPT answers and remain visually separated from the organic response. Depending on the placement, a user may see one or several advertising blocks, while an individual block can include several products or advertisers.

Chat Cards: The Core Native Format

The primary advertising format is a native Chat Card designed to fit naturally below a ChatGPT response. It includes the advertiser’s brand name and logo, a headline of up to 50 characters, a description of up to 100 characters, a square image, and a landing page URL.

Logos should be at least 128 × 128 pixels. For the main creative, OpenAI’s current API guidance recommends a square PNG or JPG image of at least 640 × 640 pixels.

The format is concise and text-led, which makes the relationship between the conversation, ad message, and landing page especially important. Advertisers can submit multiple headlines and description variations for the same offer, increasing the platform’s ability to select copy that matches a relevant conversation.

OpenAI can also generate suggested ad drafts from an advertiser’s website. AI-assisted creative production, which proposes copy and imagery based on the landing page, is currently being tested. In addition, advertisers can opt in to text customization, allowing the platform to adapt ad copy to the conversation and translate it into the user’s language.

Sponsored Agents represent another potential format, allowing users to enter a conversation with a brand’s AI agent after clicking an ad. For now, however, the product remains in a limited alpha with selected US advertisers, and OpenAI is not accepting early-access requests.

Product Feed Ads for E-Commerce and Retail

Product Feed Ads create a particularly strong opportunity for e-commerce and retail businesses. Instead of building every ad manually, advertisers can provide a catalog containing product images, titles, ratings, regular and sale prices, and brand information.

Feeds can be uploaded as CSV or TXT files, provided through a hosted URL, or transferred via SFTP. Because uploaded products expire after two weeks, a hosted URL or SFTP connection is the more practical option for maintaining an up-to-date catalog.

According to OpenAI, product feed ads are among the platform’s strongest-performing formats. As a result, they should be a priority for eligible e-commerce advertisers rather than treated as an additional creative variation.

Products can also appear in multi-item carousel cards. OpenAI reports carousel card impressions and clicks separately, giving advertisers greater visibility into how users interact with individual products.

Shopify merchants have an additional integration option through the ChatGPT Ads app by OpenAI. The app connects the store catalog, Shopify pixel, and conversion events with Ads Manager. It is currently available in the US, with international availability scheduled to begin on September 23, 2026.

Contextual Targeting Instead of Keywords

ChatGPT Ads does not use the familiar keyword targeting model found in paid search. Nor does it currently provide demographic targeting. Instead, delivery is contextual and considers the current conversation together with the advertiser’s landing page, ad copy, and optional context hints.

Context hints are added at the ad-group level and can contain up to 2,000 relevant topics or terms. They help OpenAI understand which conversations may fit the offer, but they do not work as exact-match keywords. Advertisers also cannot use them to create negative keyword lists.

This changes the planning process considerably. Rather than building campaigns around isolated search terms, advertisers need to define the problems, questions, and decision-making situations in which their offer is genuinely useful. Clear copy, focused landing pages, and specific use cases therefore play an essential role in campaign relevance.

Campaigns can run across Android app, Android web, desktop web, iOS app, and iOS web. Performance insights can be segmented by platform, helping advertisers understand where users are discovering and responding to the offer.

Custom Audiences and Bid Multipliers

Advertisers can include or exclude Custom Audiences using email addresses, phone numbers, hashed email or phone data, and Android GAIDs. Different identifier types can be combined within the same file, while lists containing more than five million records are supported.

Audience inclusion requires at least 25,000 matched users, although OpenAI recommends 100,000 or more for effective delivery. Smaller lists can still be used for exclusions. Advertisers can now add, remove, or replace audience members without rebuilding the audience from the beginning.

Audience bid multipliers provide an additional optimization layer. At the ad-group level, advertisers can apply a multiplier between 0.1x and 10x to increase or decrease the maximum bid for users included in a selected audience. If several multipliers apply, the highest one is used.

Importantly, a bid multiplier does not determine who is eligible to see the ad. It only changes how aggressively the advertiser bids when an eligible user belongs to a particular audience.

Campaign Objectives, Bidding, and Budgets

OpenAI Ads Manager currently supports three main objectives: Views, Clicks, and Conversions.

Views campaigns are billed on a CPM basis, while Clicks campaigns use CPC billing. Conversion campaigns can now operate through either optimized CPC or optimized CPM. With oCPC, the advertiser is billed for clicks, and the system optimizes toward conversions following those clicks. With oCPM, the advertiser is billed per thousand impressions, while the platform can optimize using both click-through and view-through conversion signals.

OpenAI currently recommends oCPM for conversion-focused campaigns. Each campaign can optimize toward one standard conversion event. The selected objective, billing method, and optimization event cannot be changed after the campaign is created, so advertisers need to choose them carefully. To move an existing campaign to oCPM, the campaign must be cloned with the new setup.

Maximize Results is the default automated bidding strategy and does not currently use a target CPA or target ROAS. Manual options include Max CPC and Bid Cap, while OpenAI recommends an initial maximum CPC of approximately $3–$5. Ads compete through a relevance-weighted second-price auction, meaning that a higher bid alone does not guarantee delivery if the ad is poorly aligned with the conversation.

Minimum daily budgets depend on the account currency. Current minimums include $25 in USD. A daily budget is treated as a seven-day average, so spend can reach twice the selected amount on an individual day, but should remain within seven times the daily budget over a week.

Campaign total budgets are paced across the selected period. If no end date is provided, Ads Manager uses a default 60-day pacing period. Advertisers can switch from a total campaign budget to a daily budget, but this change cannot be reversed.

Which Verticals Can Advertise?

Vertical eligibility remains one of the most important considerations before preparing a campaign setup/launch. OpenAI reviews the advertiser’s primary business, creative materials, offer, and landing-page experience rather than evaluating the ad copy in isolation.

The strongest opportunities currently include retail and e-commerce, household and consumer goods, local services, digital products and apps, education, travel and entertainment, and certain professional services. These categories align naturally with the way users rely on ChatGPT to research options, make plans, compare products, and find practical solutions.

Financial services, healthcare and medicine, and legal services are restricted to the United States. Advertisers in these categories require manual approval and must provide appropriate license documentation. The health category includes supplements, medical devices, dental services, cosmetic procedures, and health insurance. Outside the US, these regulated categories are prohibited.

Several verticals are not eligible. These include dating and adult services, AI companion products, gambling and cash-prize games, alcohol, tobacco and vaping products, cannabis, political advertising, counterfeit goods, scams, individual job or housing listings, and diet pills, detox products, or supplements promoted with unsupported claims.

Furthermore, if an advertiser’s primary business falls within a restricted category, the entire advertiser account may be considered ineligible, even if a particular campaign promotes a different offer. OpenAI may also decline advertising that conflicts with its advertising principles, business interests, or competitive position.

For Point2Web clients, the immediate focus is therefore on e-commerce and retail, product feed campaigns, digital products and apps, education, travel, local services, and consumer goods. iGaming and dating are not eligible, while nutra, financial, healthcare, legal, and other sensitive categories are not recommended for launch at this stage.

Review and Enforcement

OpenAI reviews the entire advertising journey, including advertiser verification, creatives, copy, and landing pages. Websites must be accessible to OAI-AdsBot and OAI-SearchBot, which means advertisers should avoid cloaking, restrictive geo-gating, or configurations that prevent OpenAI from evaluating the destination.

Policy violations can result in ad rejection, limited delivery, ad removal, account suspension, or termination. Severe or repeated violations may also affect related accounts belonging to the same legal entity.

For that reason, eligibility and compliance should be evaluated before campaign production begins. This prevents advertisers from investing in creatives, feeds, and landing pages that cannot pass review under the current policies.

Measurement and Performance Tracking

ChatGPT Ads already supports the core measurement capabilities required for performance campaigns. Advertisers can track website conversions through the OpenAI Pixel and Conversions API, while click-through and view-through attribution provide a broader view of how campaigns contribute to results.

For app advertisers, integrations with AppsFlyer, Adjust, and Branch make it possible to connect ad engagement with installs and in-app events. E-commerce businesses can also measure order value and ROAS, while reporting can be segmented by country, device, platform, and product.

However, reliable measurement depends on more than simply installing a tracking tool. Conversion events, attribution settings, redirects, data matching, and event deduplication must be configured correctly to avoid incomplete or duplicated results. Since reporting data may also take 24-48 hours to appear, campaigns require careful monitoring before optimization decisions are made.

We support the complete measurement process, from selecting the right conversion events and implementing tracking to validating data quality, analyzing performance, and optimizing campaigns based on reliable results. This gives clients clear visibility into campaign impact without requiring their internal teams to manage the technical setup independently.

What Early Results Tell Us

Although the platform is still developing and broad benchmarks are limited, the first published advertiser results show how strongly campaign specificity can influence performance.

Stream reported that more than 70% of its ChatGPT Ads clicks became website sessions, compared with less than 10% across other tested channels at comparable spend. The campaigns generated approximately twice the return on ad spend and were the only tested channel to produce both sign-ups and a qualified pipeline.

Jotform found that more than 10% of ChatGPT Ads clicks converted into sign-ups. Its cost per sign-up was approximately 30% lower than paid search and 60% lower than paid social. Use-case-specific landing pages increased the sign-up rate by 2.5 times, while moving from Reach to Click campaigns reduced CPC by four to five times.

Meanwhile, Newegg’s product feed campaigns connected users with relevant products while they were actively researching what to buy, supporting OpenAI’s position that feed ads are particularly promising for e-commerce. LegalNature also found that campaigns built around individual document types performed better than broad legal-service messaging, highlighting the importance of aligning each offer with a specific user need.

These results should not be treated as universal benchmarks, since they come from individual advertisers, campaign periods, and attribution methodologies. Still, they reinforce an important principle: ChatGPT Ads appear most effective when the campaign, ad, landing page, and user’s immediate task are closely connected.

Launching ChatGPT Ads With Point2Web

ChatGPT Ads is still an emerging channel, which means advertisers should approach it with a structured testing strategy. Formats, policies, inventory, and optimization capabilities will continue to develop, while performance will vary considerably by vertical, offer, market, and measurement quality.

We now provide clients with the infrastructure and full support needed to work with ChatGPT Ads. Our team can help assess advertiser and vertical eligibility, prepare the account structure, develop contextual campaign strategies, create and test ads, connect product feeds, implement conversion tracking, manage bidding and budgets, and optimize performance based on reliable data.

For eligible advertisers, particularly in e-commerce, digital products, apps, education, travel, local services, and consumer goods, the current stage offers an opportunity to begin learning before conversational advertising becomes a more established part of the media mix.

Interested in testing ChatGPT Ads for your product? Contact us to get your account, target markets, campaign format, and measurement setup, and build a launch strategy around the opportunities available today.