ChatGPT Ads partner for advertisers preparing for OpenAI's conversational ad platform across five markets.

It depends on where you are and what you sell. Self-serve advertising at ads.openai.com launched in the United States on 5 May 2026, with no minimum spend and both CPC and CPM bidding. UK, Australia, New Zealand, and Canada have access through managed and agency channels. Singapore and Malaysia are not live yet. Four verticals are excluded regardless of market: financial services, health, dating, and politics. If you are in an eligible market with an eligible product category, yes you can start today. If you are elsewhere or in an excluded vertical, the right move is organic GEO to build citation presence while the platform expands. Source: OpenAI Help Center article 20001047.

As of August 2026: the United States has self-serve access via ads.openai.com (launched 5 May 2026). The United Kingdom, Canada, Australia, and New Zealand have managed and agency access. OpenAI has not published a confirmed timeline for further market expansion. All other markets, including Singapore and Malaysia, are not live. Source: OpenAI Help Center article 20001047.

No. As of August 2026, neither Singapore nor Malaysia has live ChatGPT Ads access (self-serve or managed). OpenAI has not published a confirmed rollout date for either market. The practical path for Singapore and Malaysia advertisers right now is organic GEO: building your brand's citation presence inside ChatGPT responses through structured content, authoritative sourcing, and AI-native publishing. That foundation will matter regardless of when paid access arrives.

No. Financial services is one of four verticals excluded from ChatGPT Ads by OpenAI policy. Health, dating, and politics are also excluded. This covers insurance, banking, wealth management, lending, investment products, and related regulated sectors. The exclusions apply across all markets where ChatGPT Ads is available. For financial services and insurance brands, the right strategy is organic GEO: earning citation in ChatGPT responses through authoritative content, structured data, and regulatory-compliant positioning. We run GEO programmes for excluded-sector clients across Singapore, Malaysia, Australia, Canada, and the US. See also our pages on insurance marketing and banking-finance visibility. Source: OpenAI Help Center article 20001047.

Today the live format is Sponsored Answers: a labelled ad unit appearing within ChatGPT responses, visually distinct from the answer itself. Users can trigger the Ask ChatGPT mechanic (three-dot menu on the ad) to share the ad in-chat; ChatGPT then answers about the product independently and without influence from the ad content. Forthcoming formats confirmed by David Dugan, OpenAI head of advertising, in Digiday interviews on 30 June and 1 July 2026: image ads, video ads, conversational ads, native formats, and interactive ads. OpenAI is hiring San Francisco engineers to build these. No launch dates have been published for the forthcoming formats.

No, and OpenAI has stated this explicitly. The ad system and the answer system are separate. Advertising on ChatGPT does not influence how ChatGPT responds to user queries. The Ask ChatGPT mechanic works the other way: a user who sees an ad can click the three-dot menu, share the ad in-chat, and ask ChatGPT about the product. ChatGPT then answers based on its training and retrieval, not based on advertiser instructions. This separation is by design. Source: OpenAI Help Center article 20001047.

Available: aggregated views, clicks, conversion pixel data, Conversions API integration, attribution windows, and CPC plus CPM bidding signals. Not available: individual chat content, user memories, names, email addresses, precise location, or IP addresses. Retention: data is held for 30 days post-deletion. Third-party measurement is not yet formally supported, but David Dugan called it a natural next step in Digiday, 30 June 2026. The data picture is narrower than Google or Meta. Plan measurement around aggregated signals and incrementality holdouts rather than individual-level attribution. Source: OpenAI Help Center article 20001047.

CPC (cost per click) and CPM (cost per thousand impressions) bidding are both available. There is no minimum spend requirement for self-serve accounts. OpenAI has not published benchmark CPC or CPM figures, so there is no credible published rate card to reference. We do not cite invented cost figures. Actual rates depend on category, audience, and competitive demand in the auction. We model this from account data once campaigns are live.

A Sponsored Answer is the current live ChatGPT ad format. It appears within the ChatGPT interface as a labelled unit, visually distinct from the conversational response. The label makes the commercial nature of the placement transparent to users. The ad unit itself does not alter the text of ChatGPT's answer to the user's query; the systems are separate. The Ask ChatGPT mechanic lets a user who sees the ad engage further by sharing it into the chat, prompting an independent ChatGPT response about the advertised product. Source: OpenAI Help Center article 20001047.

A display ad interrupts. A conversational ad participates. The distinction matters because ChatGPT users are in active reasoning mode, not passive browsing mode. A conversational ad format (announced by David Dugan in Digiday, 30 June 2026, not yet live) would sit inside the dialogue flow rather than alongside it. The creative brief is different: it needs to read like an answer, not like a banner. The measurement challenge is different too: dwell, follow-on queries, and in-chat engagement matter more than click-through rate. We are building creative and measurement frameworks for conversational formats ahead of their launch.

David Dugan is OpenAI's head of advertising. In Digiday interviews on 30 June and 1 July 2026, he confirmed that OpenAI is developing image, video, conversational, native, and interactive ad formats, and is hiring San Francisco engineers to build them. He also described third-party measurement as a natural next step for the platform. These statements are the most authoritative published signal on where ChatGPT Ads is headed. No launch timelines were attached to the forthcoming formats.

Three structural differences. First, individual-level signals are absent: no individual chat content, no user-level identity signals, no precise location or IP. Google and Meta both carry far richer individual signals. Second, no third-party verification yet: Google supports IAS and DoubleVerify; Meta supports multiple MRC-accredited vendors. ChatGPT has not yet formalised third-party measurement access, though Dugan cited it as a natural next step (Digiday, 30 June 2026). Third, conversational context is unmeasured: an ad seen in the middle of a research conversation may influence purchase intent differently than a feed impression. Standard view-through and click-through metrics capture only part of the picture. Build your measurement plan around aggregated signals, conversion pixel, Conversions API, and incrementality holdout design.

It is not formally available yet. OpenAI has not announced a third-party verification programme. David Dugan described it as a natural next step (Digiday, 30 June 2026). For now, plan around OpenAI's aggregated reporting, your own conversion pixel and Conversions API data, and incrementality holdout tests. If brand-safety verification is a contractual requirement in your organisation, that is a real constraint that may delay ChatGPT Ads activation until the platform matures.

For US advertisers preparing for a ChatGPT Ads launch, the 90 days before launch checklist: (1) Category review against OpenAI excluded-category list. (2) First-party data audit and Conversions API readiness. (3) GEO baseline measurement so you know where you stand organically before paid begins. (4) Sponsored Answer creative brief and copy tested internally. (5) Attribution window and incrementality holdout design, so measurement infrastructure is ready from day one. (6) 2027 budget line proposal for board or finance team approval. The structural decisions made before launch day compound; getting measurement architecture right at the start is worth more than optimizing it after the fact.

The platform is new, the measurement environment is thin, and there is no published performance benchmark. Running it yourself means learning the hard way on live budget. The agency value on an emerging platform is different from a mature one: it is less about execution efficiency and more about structural readiness. Getting measurement right from day one, understanding what data is and is not available, writing creative that works in a conversational context rather than a feed context, and connecting ChatGPT Ads into a broader cross-engine visibility strategy. Those decisions compound. Getting them wrong in month one costs more than the management fee.

A senior practitioner from the leadership team is on every account. Strategic direction set by Siddharth Surana (Founder and CEO, 18+ years), Sundeep Surana (Managing Director, 16+ years), Ratnakar Nemani (Operations Director, 11+ years), or Nitesh Sanghvi (Search and Social Director, 12+ years). Team based in Singapore and India. One team, two time zones, no white-labelling, no junior-managed accounts.

Yes, when the incumbent arrangement is up. Handover: 1 to 2 week diagnostic on account state (campaign structure, Sponsored Answers inventory, conversion pixel health, Conversions API setup, attribution window calibration), formal handover of ads.openai.com access, parallel-run period if needed, full operational responsibility at the agreed date.

Written findings covering: category eligibility (your vertical against OpenAI's excluded-category list), market availability (your target countries against live self-serve and managed markets), first-party data infrastructure readiness, GEO baseline (where you currently rank in ChatGPT responses organically), conversion pixel and Conversions API setup requirements, creative brief for Sponsored Answers, cross-engine visibility gap analysis, and a 90-day activation plan. 2 to 3 weeks. Fixed scope.

Delivery includes a payback period estimate for the test budget, based on your category, target markets, and current measurement infrastructure. Finance teams often want to see the payback math before committing even a test line; we build that in.

Organic GEO: generative engine optimization. The goal is to be the source ChatGPT cites or draws on when users ask questions in your category. This involves structured content, authoritative sourcing, AI-native publishing formats, and Speakable schema. For financial services and insurance, where MAS FAA-N03 and related regulatory frameworks govern what you can publish, GEO content must be compliance-reviewed before publication. We run GEO programmes for excluded-sector clients and cross-link with our service pages for insurance, banking-finance, and fintech. GEO compounds; the organic presence you build now will matter when and if paid access opens for your category.

They operate on different surfaces but reinforce each other. Paid Sponsored Answers appear in the interface as labelled ad units. Organic GEO influences the unpaid, unmarked portion of ChatGPT responses. A user who sees a Sponsored Answer and then asks a follow-up question about your category will get an answer shaped by your organic GEO presence. The brands that win on conversational platforms are those that have both: visibility in paid placements and authority in organic retrieval. Running ChatGPT Ads without an organic GEO programme means the paid impression leads to an organic response the brand does not control.

For categories and markets where it is live: yes, as a dedicated budget line within your test allocation, not a primary channel. ChatGPT claims approximately 1 billion weekly active users (Sam Altman public statement, early 2025). The reach is smaller than Google and Meta. The measurement environment is thinner. But the opportunity window for early learners is real; platform habits form faster on channels where the early movers are testing rather than scaling.

Plan a test budget, establish your measurement baseline, and build the creative and data infrastructure now. For excluded categories: budget for GEO, not ChatGPT Ads, until OpenAI's exclusion list changes. For markets where it is not live (Singapore, Malaysia): build organic GEO foundation now, allocate a contingency line for 2027 if market expansion accelerates. When your board asks about the next marketing dollar on AI-native platforms, ChatGPT Ads belongs in that conversation alongside Google AIO and Perplexity.

Potentially yes for creative generated using AI tools. EU AI Act Article 50 (live from August 2026) requires disclosure when AI-generated content, including text and image, could be mistaken for human-created material. California SB 942 (effective January 2026) requires AI-generated content disclosures in commercial contexts. For ChatGPT ad creative, if copy or imagery is generated via AI tools, disclosure requirements may apply depending on jurisdiction and the nature of the content. The clearer obligation is OpenAI's own ad policies, which apply uniformly to all advertisers on the platform. We review creative against applicable disclosure requirements at the brief stage.

We are not pretending to have three years of ChatGPT Ads case studies. The channel launched in May 2026. What we have: the GEO and AEO measurement stack that the platform assumes, built and running before the paid channel launched. Senior practitioners who have navigated platform launches before, from programmatic in 2011 to Performance Max in 2021 to AI Max in 2025. A measurement discipline that does not rely on the platform's own attribution as the single source of truth. And a posture that tells you what the platform cannot do yet, not what you want to hear. For a channel this early, that is the relevant credential.