Updated 6 Oct 2026: Adds OpenAI's Oct 5 visual ad test during ChatGPT image generation plus expanded conversion-data/attribution partners, geo-based incrementality work, brand-suitability pilots and selected early partner-reported performance evidence.

Key details

  1. OpenAI announced Sponsored Agents on September 16, 2026.
  2. Sponsored Agents are currently a limited alpha with selected advertisers in the United States.
  3. Users can enter a clearly labelled business-sponsored AI conversation after clicking a ChatGPT ad.
  4. OpenAI says sponsored conversations are separate from ChatGPT's independent answers and the original conversation.
  5. Ads Manager gains additional AI-assisted campaign, creative and text-customisation workflows.
  6. HubSpot is OpenAI's first CRM partner for ChatGPT Ads and Shopify its first ecommerce partner.
  7. Amazon DSP remains a separate pilot buying path into ChatGPT Ads for selected U.S. advertisers.
  8. On September 24, ChatGPT Ads began expanding to seven additional Asian markets, taking availability to more than 60 countries.
  9. OpenAI says ChatGPT Ads passed a $1 billion annualized revenue run rate by the end of August and now has tens of thousands of advertisers; these are company-reported figures.
  10. On September 30, Ads Manager added spreadsheet bulk creation for product-feed campaigns/ad groups and product review-status visibility.
  11. The public Insights API can now report attributed conversion events beyond the campaign's selected optimization goal without changing that goal.
  12. Account, public brand and legal billing names can now be managed separately; brand/legal changes can trigger review or pause serving.
  13. On October 5 OpenAI announced a visual ChatGPT Ads format to be tested later in October during image generation with an initial group of U.S. advertisers.
  14. OpenAI added conversion-data integrations with Hightouch, Tealium and LiveRamp and expanded its attribution/measurement partner ecosystem.
  15. OpenAI is exploring geo-based incrementality measurement with Haus, Measured and WorkMagic.
  16. DoubleVerify and Integral Ad Science are participating in brand-suitability evaluation pilots.
  17. OpenAI published selected partner-reported early performance results; these are not platform-wide independent benchmarks.

What builders should take away

  1. Treat Sponsored Agent engagement as a distinct funnel stage rather than collapsing it into an ordinary ad click.
  2. Instrument backend revenue and lead outcomes so richer in-ChatGPT engagement does not become a vanity conversion metric.
  3. Define what claims, recommendations and handoffs a sponsored agent may make before testing the format.
  4. Keep human review around AI-generated campaign creative and material campaign changes.
  5. If using HubSpot or Shopify integrations, document which system is authoritative for lead, catalogue, conversion and campaign state.
  6. Compare direct OpenAI buying and Amazon DSP access on operational control and incrementality rather than platform-reported attribution alone.
  7. For product-feed advertisers, move repeat campaign creation into the bulk-sheet workflow and use product review status to diagnose rejected catalogue items.
  8. Use the broader Insights API conversion reporting to inspect downstream outcomes without changing the event the campaign is optimizing for.
  9. Separate platform-attributed conversions from incrementality: use the new partner ecosystem to test whether ChatGPT Ads create additional demand rather than merely claiming existing conversions.
  10. For visual inventory, review whether image-generation context changes creative, suitability and landing-page assumptions compared with ordinary conversational placements.

What changed

On September 16, OpenAI announced a limited U.S. alpha of Sponsored Agents. After clicking an ad in ChatGPT, eligible users can choose to start a separate, clearly labelled conversation with an AI representative sponsored by that business, ask follow-up questions and then follow a link to the advertiser's website. OpenAI says the sponsored conversation is distinct from ChatGPT's independent answers and separate from the user's original conversation. The same release adds prompt-driven campaign creation in ChatGPT Work, new Ads Manager creative tools, AI text customisation, and integrations with HubSpot and Shopify as OpenAI's first CRM and ecommerce partners. This follows the platform's self-service global expansion and Amazon DSP pilot already covered here. On September 24, OpenAI began rolling ChatGPT Ads out in Indonesia, Malaysia, the Philippines, Singapore, Thailand, Vietnam and Taiwan, taking advertiser availability to more than 60 countries. OpenAI says the ad business passed a $1 billion annualized revenue run rate by the end of August and now has tens of thousands of advertisers. On September 30, an advertiser update added spreadsheet bulk creation for product-feed campaigns and ad groups, product review-status visibility, Insights API reporting for attributed conversion events beyond the campaign's selected optimization goal, and separate account, public brand and legal-name controls. On October 5, OpenAI announced a new visual ad format that it plans to test later in October with an initial group of U.S. advertisers during image generation in ChatGPT. OpenAI also expanded the measurement stack with conversion-data integrations for Hightouch, Tealium and LiveRamp; attribution/reporting partnerships across web and app measurement providers; early geo-based incrementality experiments; and brand-suitability pilots with DoubleVerify and Integral Ad Science.

Why it matters

The advertising unit is becoming an interactive conversion surface rather than only a placement. A click can now lead into a branded AI sales or product-discovery conversation inside ChatGPT before the user reaches the advertiser's site. For marketers and builders, that changes funnel design, measurement and governance: some qualification and persuasion can happen inside a sponsored agent, while campaign operations and creative production are also becoming AI-assisted. The alpha boundary matters, however: Sponsored Agents are not generally available and OpenAI has not published a broad rollout date. The October 5 release pushes the platform beyond campaign operations into richer creative inventory and more serious causal/independent measurement. Ads appearing during image generation create a new contextual surface, while incrementality and brand-suitability work address two gaps that ordinary platform attribution cannot resolve.

Sponsored Agents add a post-click conversational layer

Selected U.S. advertisers can test a format where a user clicks a ChatGPT ad and optionally opens a clearly labelled conversation with a business-sponsored AI representative. OpenAI says this conversation remains separate from its independent answers and from the original user conversation.

The format is still a limited alpha

OpenAI's Help Center says Sponsored Agents are available only to selected advertisers in a limited alpha test. General early-access requests are not currently open, so this is an emerging format rather than standard ChatGPT Ads inventory.

Ads Manager is getting more AI-native campaign operations

OpenAI says advertisers can use natural-language prompts to create, update and analyse campaigns, generate or recommend creative from landing-page context, and use AI text customisation to adapt copy and translations to conversation context.

HubSpot and Shopify move ChatGPT Ads into existing business systems

OpenAI describes HubSpot as its first CRM partner and Shopify as its first ecommerce partner for ChatGPT Ads. These integrations reduce the need to treat the ad platform as a standalone workflow and create new paths between campaigns, product catalogues and lead follow-up.

Amazon DSP remains a parallel buying path

The earlier Amazon Ads pilot still gives selected U.S. advertisers another route into ChatGPT inventory through Amazon DSP. Sponsored Agents add a new interaction format on top of the broader move toward a multi-entry advertising ecosystem.

Measurement now has another boundary to instrument

Teams testing Sponsored Agents will need to distinguish ad impression, click, sponsored-agent engagement, qualified conversation, site visit and backend conversion. Platform engagement can become richer without necessarily proving incremental revenue.

Product-feed operations and measurement get less manual

The September 30 advertiser update adds spreadsheet bulk creation for product-feed campaigns and ad groups, product-level review status, and Insights API access to attributed conversion events beyond the campaign's selected optimization goal. Advertisers can also separate internal account, public brand and legal billing names; brand-name changes trigger review and legal-name changes restart business review.

Visual ads move into image-generation sessions

OpenAI says it will begin testing a new visual ad format later in October with an initial group of U.S. advertisers during image generation in ChatGPT. The ads are clearly labelled and separate from the generated image and from ChatGPT's answers.

Measurement expands beyond OpenAI's own reporting

OpenAI added conversion-data integrations with Hightouch, Tealium and LiveRamp and named a broad group of web/app attribution partners. It is also working with Haus, Measured and WorkMagic on geo-based incrementality experiments and with DoubleVerify and Integral Ad Science on brand-suitability evaluation pilots.

Early performance numbers are partner-reported, not yet a platform benchmark

OpenAI cites early partner findings: DV Rockerbox reported WeightWatchers' attributed CPA 15.3% below its blended paid-search benchmark; WorkMagic reported statistically significant lift for Dose with 67% of incremental purchases from new customers; and Triple Whale reported 93% of Portland Leather visitors from ChatGPT Ads were new. These are selected partner/customer results rather than independent platform-wide performance evidence.

What to watch next

  • When Sponsored Agents expand beyond the limited U.S. alpha.
  • Eligibility, pricing and minimum-spend requirements for Sponsored Agents.
  • What advertiser data and tools sponsored agents can access during conversations.
  • Reporting for sponsored-agent starts, engagement depth, site handoffs and conversions.
  • Broader HubSpot and Shopify integration availability and API surfaces.
  • Whether other CRM, ecommerce and DSP partners gain comparable ChatGPT Ads integrations.
  • Whether the broader international rollout changes Sponsored Agents eligibility beyond the current selected-U.S.-advertiser alpha.
  • Whether bulk product-feed creation and broader conversion reporting gain corresponding automation/API write surfaces beyond Ads Manager and the existing Insights API.
  • Results and broader availability of the visual image-generation ad test.
  • Whether geo-based incrementality and DoubleVerify/IAS brand-suitability pilots become generally available advertiser products.
  • Independent or larger-sample evidence on ChatGPT Ads CPA, new-customer lift and incremental revenue.

Still unclear

  • Sponsored Agents are a limited alpha and OpenAI has not announced general availability.
  • OpenAI has not yet published enough public detail to establish the full pricing or measurement model for Sponsored Agents.
  • The practical conversion lift of sponsored conversations has not been independently established.
  • Advertiser workflow and integration capabilities may change substantially during alpha testing.
  • The visual format is an upcoming U.S. test rather than generally available inventory.
  • The published performance examples are selected partner-reported customer results and do not establish typical ChatGPT Ads performance.

Sources

Direct reading behind this dossier.

8 sources
Reimagining advertising with AI
OpenAI primary

Primary announcement for Sponsored Agents, AI-assisted Ads Manager changes, HubSpot and Shopify integrations.

Building advertising for the way people use AI
OpenAI primary

Primary announcement for visual ads during image generation, expanded measurement integrations, incrementality work, brand-suitability pilots and selected early partner findings.

Discussion

Discussion is reader-contributed. Comments are not part of the BTN dossier or its editorial evidence.

0 visible comments

Join the discussion

Keep comments useful and relevant. Reader contributions may be moderated and are not BTN editorial evidence.

Sign in to comment