ChatGPT Ads in Singapore: How They Work and Which Brands Should Test Them
ChatGPT Ads are now rolling out in Singapore and five other Southeast Asian markets.
Eligible businesses can use OpenAI’s self-service Ads Manager to reach people during relevant conversations. For brands whose customers research, compare or ask for advice before buying, this creates another opportunity to enter the consideration process. Early advertiser results are encouraging, but they vary considerably. The sensible starting point is a focused test with clear commercial targets, alongside your existing search, social and commerce campaigns.
Here is how the channel works, what earlier advertisers have reported, and how Singapore and regional brands can decide whether it deserves a place in their media plan.
Table of Contents
What are ChatGPT Ads, and how do they work?
ChatGPT Ads are paid placements displayed separately from ChatGPT’s answers. An ad can include a brand name, headline, copy, image and a link to an advertiser’s landing page. Paying for an ad does not buy influence over the answer itself.
OpenAI uses conversational context to help determine relevance. A person researching a laptop for video editing, for example, presents a different opportunity from someone asking how to repair their current computer. Advertisers should build their messaging around the problem their product solves.
This differs from buying a specific search keyword. Context hints can help describe suitable situations, but they do not guarantee delivery against an exact prompt. Advertisers also do not receive users’ private conversations.
Ads currently reach eligible adult users on Free and Go plans. Paid plans like Plus and Pro, remain ad-free. As we will discuss later, this will affect which businesses ChatGPT Ads are more suitable for.
What types of ChatGPT Ads campaigns can brands run?
The clearest way to understand the options is to separate campaign objectives from how the ads are created. Views, Clicks and Conversions describe what delivery is optimised to achieve. Manually created ads and product feed ads describe how the creative and product information are supplied.
| Objective | What it prioritises | How you are billed | How we would evaluate it |
|---|---|---|---|
| Views | Visibility and exposure. | Per 1,000 impressions (CPM). | Delivery cost, followed by evidence of meaningful brand or business impact. |
| Clicks | Visits to your website or product destination. | Per valid click (CPC). | CPC, landing-page sessions and the quality of subsequent actions. |
| Conversions | A selected supported action, such as a purchase, sign-up or lead submission. | Per valid click (oCPC) or per 1,000 impressions (oCPM), depending on the selected billing option. | Conversion volume, cost per acquisition, revenue and customer quality. |
A conversion objective requires conversion tracking through OpenAI’s JavaScript Pixel and/or Conversions API, plus an available supported standard event. It does not mean you pay only when someone converts. Click billing optimises towards conversions after an ad click; impression billing can account for eligible conversions following both views and clicks.
Source: OpenAI’s conversion-optimised campaign guide.
Our recommendation is to choose the objective from the business question. A traffic test can help assess whether visitors engage with a useful comparison page. A sales test should work towards purchases when tracking and event volume support it. A visibility campaign needs its own success criteria: impressions alone cannot show whether consideration improved.
Two ways to build ads: manual creative and product feeds
Manual Creative
Once you have chosen your campaign objective, the next decision is how to build the ads. With manually created ads, you develop the message around a particular offer and send people to a relevant landing page. This approach suits businesses promoting a service, a subscription trial or a focused seasonal offer, where the copy needs to explain why someone should take the next step.
Product Feed
For retailers with a larger catalogue, product feeds provide a way to create ads across multiple items using structured information such as product titles, descriptions, images, prices, availability and destination URLs. Instead of producing every ad individually, the retailer can use templates and select product groups to advertise. A computer retailer, for example, might organise its catalogue into gaming laptops, monitors and components, with each group supported by details that help shoppers assess suitability.
The quality of that catalogue matters because it shapes what customers see and what they find after clicking. Useful specifications make products easier to compare, while current prices and stock information help ensure the offer matches the purchase experience. Maintaining the feed should therefore be part of campaign management, alongside reviewing creative and performance.
Manual creative and product feeds describe how ads are built; the campaign objective determines what delivery works towards. The available combinations depend on your account and setup, so check these when planning a campaign. Product feeds used for advertising also remain separate from organic discovery: uploading a feed does not automatically place your products in ChatGPT’s answers.
Sources: OpenAI’s product feed campaign guide and product feed documentation.
Why Southeast Asian marketers should pay attention
On 23 September 2026, OpenAI announced expansion into Indonesia, Malaysia, the Philippines, Singapore, Thailand and Vietnam, alongside Taiwan. These are six Southeast Asian markets, with Taiwan included separately in the same announcement.
ChatGPT was already part of many consumers’ digital routines before advertising launched in the region. A comparable breakdown from GWI’s Q2 2025 survey, published in the Digital 2026 Global Overview Report, shows how adoption differed across the six Southeast Asian launch markets.
| Market | Monthly ChatGPT usage |
|---|---|
| Malaysia | 48.4% |
| Philippines | 42.4% |
| Indonesia | 37.9% |
| Vietnam | 37.3% |
| Singapore | 37.1% |
| Thailand | 23.3% |
Measure: self-reported use of ChatGPT through its website or app within the past month. Survey period: Q2 2025; report published October 2025. These are historical usage rates among internet users aged 16+, not shares of the total population or estimates of advertising reach.
Malaysia had the highest reported usage among these six markets, with almost half of surveyed internet users using ChatGPT each month, followed by the Philippines. Indonesia, Vietnam and Singapore were closely grouped at around 37–38%, while Thailand recorded a lower rate. Five of the six markets were above the worldwide benchmark, suggesting that ChatGPT had already established a substantial presence across much of the region before the advertising rollout.
Source: GWI’s ChatGPT usage chart, reproduced by Meltwater from the Digital 2026 report.
More recent evidence also points to strong adoption among younger users. DataReportal’s April 2026 mid-year report puts monthly ChatGPT usage among internet users aged 16–24 at 67.4% in Malaysia and 66.8% in the Philippines. Those figures describe a narrower age group and should be read separately from the country-wide baseline above.
Source: DataReportal’s Digital 2026 Mid-Year report.
For media planning, usage is a starting point rather than a budget allocation rule. A higher adoption rate does not establish stronger purchase intent, cheaper acquisition or a larger absolute audience, and the survey includes users who may not be eligible to see ads. We would use this baseline to assess where customer research may already be taking place, then validate the opportunity against the brand’s audience, margins and actual campaign delivery.
That validation also needs a local purchase experience. In Singapore, clear SGD pricing, delivery information, warranties and credible reviews can help remove uncertainty; elsewhere, language, payment preferences and product availability need to reflect the market being served. A regional campaign becomes more useful when the offer and landing page answer local customers’ questions, rather than simply extending the same message across six countries.
Have ChatGPT Ads worked in other markets?
There are successful examples from advertisers who started earlier. Their results provide useful hypotheses for a Singapore test, rather than a reliable forecast of local performance. The customer stories below were published by OpenAI and draw on advertiser reporting; they are selected examples, not independent channel-wide benchmarks. They also do not provide a complete country breakdown.
Newegg: evergreen product feeds plus a seasonal campaign
Newegg reported 3× return on ad spend across its ChatGPT campaigns over 28 days. A dedicated two-week Fantastech Sale campaign achieved 7× ROAS.
The distinction matters. Newegg used an evergreen product feed approach alongside a separate seasonal CPC campaign. The 7× figure belongs to the sale campaign; it should not be presented as the result of all Newegg advertising or of product feeds alone.
Computer shopping is a useful example of the channel’s potential. Buyers often need to compare specifications, compatibility, budgets and use cases before choosing. A relevant product offer can help at that decision point.
The lesson for retailers is to combine a useful catalogue with clear purchase reasons. A strong sale, current stock and a relevant product page can matter as much as the placement.
Source: OpenAI’s Newegg customer story.
Talkpal: connecting advice to a trial
Language-learning platform Talkpal reported that ChatGPT became one of its top three sources of new trials within two months. It also reported approximately 26% higher ROAS than one established native advertising channel, measured from 23 June to 18 August 2026, and 51% lower cost per trial activation against that same channel from 1 July to 9 September 2026.
Those are separate comparisons with different reporting periods. The 26% figure is not a comparison against every advertising channel, and trial acquisition is not the same as paying-customer acquisition.
For subscription businesses, the useful lesson is to measure what happens after the trial: activation, paid conversion and retained revenue.
Source: OpenAI’s Talkpal customer story.
Stream: reaching people solving a technical problem
Stream, which provides chat and video infrastructure for developers, reported roughly 2× ROAS in the test described by OpenAI. Its products address specific implementation problems, making this a useful example for specialist B2B brands.
A person working out how to add chat to an application may be open to a relevant tool. A broad corporate message is less useful than a clear explanation of what the product does and how to start.
Source: OpenAI’s Stream customer story.
But other results have also been disappointing
Agency DAC reported a client test in which the cost of booked leads was around 100 times the client’s usual range. It also described substantially higher CPMs in two tests. The published account does not disclose enough detail to generalise across sectors or markets.
Together, these examples point to a practical conclusion: conversational relevance can create an opportunity, but it does not guarantee affordable acquisition. Track the business outcome and compare it with the next best use of the budget.
Source: DAC’s assessment of ChatGPT Ads.
Which products and clients are a good fit?
Our starting hypothesis is that ChatGPT Ads suit products people need help choosing. Evaluate the buying decision rather than assuming an entire industry will perform well.
| Potential fit | Why a conversation could matter | A useful destination |
|---|---|---|
| Electronics and specialist retail | Buyers compare specifications, compatibility and budgets. | A product page or comparison collection. |
| Home, lifestyle and pet products | Customers research suitability for particular spaces or needs. | A page explaining dimensions, materials and use cases. |
| Learning and subscription products | Users ask how to develop a skill or solve a recurring problem. | A trial page with a clear explanation of value. |
| B2B software and developer tools | Buyers describe a workflow or technical requirement. | A relevant solution page, demo or documentation. |
| Considered services | Customers compare providers, scope, pricing and process. | A focused service page with evidence and a clear next step. |
A useful screening question is: What could a prospective customer ask that would make our offer helpful?
For example, a pet brand might identify questions about washable beds for small flats. A generic “premium pet products” message says little about that need. A page explaining washable covers, sizing and delivery makes the offer easier to assess.
Brands with thin margins, weak landing pages or no reliable conversion tracking should address those constraints first. Even relevant traffic becomes expensive when the purchase experience fails.
How should ChatGPT Ads work alongside other digital media?
Give each channel a clear role in the customer’s decision. ChatGPT can contribute during research and consideration, while other placements create awareness, capture demand or bring customers back.
| Channel | Role in a coordinated plan |
|---|---|
| Meta, TikTok and YouTube | Demonstrate the product, build familiarity and introduce a need. |
| ChatGPT Ads | Present a relevant offer when someone is exploring a problem or comparing options. |
| Google Search and Shopping | Capture explicit searches for products, brands and purchase options. |
| Display and video placements | Extend awareness and support permitted audience strategies. |
| Shopee and other marketplaces | Convert demand where customers prefer to transact. |
| Email and CRM | Support follow-up, repeat purchases and retention. |
Consider an illustrative laptop campaign. Creator videos demonstrate editing performance. A ChatGPT placement directs relevant researchers to a laptop comparison page. Search campaigns capture later brand or model searches. Website audiences can support follow-up on other platforms where available and permitted, while CRM helps customers after purchase.
That is a planning example, not a reported MajorForm campaign. It does not assume you can identify people from their private ChatGPT conversations or retarget someone simply because they entered a particular prompt.
Keep the offer consistent across channels, but adapt the creative to each placement. Our digital marketing funnel guide explains how to plan around customer awareness and intent.
What do ChatGPT Ads cost? Published CPC and CPM examples
There is no dependable public Singapore-wide CPC or CPM benchmark yet. Published figures come from particular campaigns, with different objectives and audiences.
Agency 2POINT shared the following results from two tests:
| Reported test | Spend | Impressions | Clicks | CPC | Calculated effective CPM | Reported leads |
|---|---|---|---|---|---|---|
| B2B | $2,001 | 32,016 | 381 | $5.25 | $62.50 | 5 |
| Home services | $196 | 3,498 | 60 | $3.27 | $56.03 | 3 |
The source uses dollar figures without explicitly specifying the currency; they are reproduced as published and are not Singapore-dollar estimates. Effective CPM is our calculation: spend ÷ impressions × 1,000. It is a delivered cost, not a quoted rate card.
The corresponding cost per reported lead is about $400 and $65. These small samples demonstrate why CPC alone is insufficient: the commercial outcome depends on how many clicks become useful customers.
Source: 2POINT’s published campaign results.
OpenAI supports impression-based and click-based buying. Conversion optimisation can also use click or impression billing; choosing a conversion objective does not mean you pay only when a conversion happens. A maximum CPC bid, where used, is a ceiling rather than an average paid price.
Source: OpenAI’s campaign creation guidance.
Set your budget from your economics. As an illustrative calculation, if you can afford S$60 per new customer and expect 5% of paid clicks to become customers, your starting allowable CPC is S$3. That assumption needs validation; it is not a forecast of ChatGPT performance.
How to measure whether the campaign is effective
Build reporting around three questions: did the campaign deliver, did visitors take useful actions, and did those actions create value? Use UTMs, site analytics and supported conversion tracking. Connect leads to CRM outcomes so a low-cost enquiry is not counted as a success when it never becomes qualified pipeline. For subscriptions, follow trial cohorts through activation and payment. For retail, distinguish new customers from existing customers and consider returns and margin.
Before comparing channels, align the conversion definition, attribution window, currency and revenue treatment. A platform’s attributed ROAS and your analytics tool’s last-click ROAS can tell different stories.
Attribution is not proof of incremental growth. An attributed order may have happened anyway, or another channel may have helped create the demand. Where volume allows, use holdouts or controlled experiments. Smaller advertisers can start with a documented baseline and a staged test, while recognising that promotions and seasonality make causal conclusions harder.
Choose one primary success metric before launch. Then set a spend limit, a review date and a minimum amount of evidence needed to scale. A good CTR can justify investigating the landing page; it cannot substitute for profitable customer acquisition.
How could ChatGPT Ads develop next?
OpenAI has already announced developments including AI-assisted ad creation, commerce integrations and sponsored agents. Sponsored agents began as a test with selected US advertisers, allowing a separate sponsored interaction with a business. That announcement should not be read as general Singapore availability.
Source: OpenAI’s advertising product update.
Our expectation is that three areas will become more important:
Better catalogue and commerce connections. Accurate product data could help brands connect detailed research to a suitable purchase destination.
More useful localisation. Regional campaigns will benefit from matching language, availability and purchase expectations to each market.
Stronger outcome measurement. Advertisers will need clearer evidence of customer quality and incremental value as budgets grow.
These are strategic expectations, not promises about OpenAI’s roadmap. More advertiser competition could also raise acquisition costs. Being early is valuable when it produces useful learning; it is not a guarantee of permanently cheap traffic.
One Last Note: Keep SEO, GEO and AEO working alongside paid media
Search engine optimisation, generative engine optimisation and answer engine optimisation help make your brand’s information accessible and useful during research. They complement paid placements, while influencing different routes to discovery.
For your own site, publish clear answers to real customer questions. Use descriptive headings, factual comparisons, original evidence and consistent product information. Make pages crawlable and maintain relevant structured data that matches visible content. Keep brand, delivery and service details consistent across your site and credible third-party listings.
Monitor traditional search visibility, referrals from AI services and the accuracy of sampled AI answers. Treat these as separate measures from paid campaign performance. No content format or schema guarantees a ChatGPT recommendation, and buying ads does not buy organic visibility.
Frequently Asked Questions (FAQ): ChatGPT Ads
1. What industries work best for TikTok Live Selling?
Beauty, fashion, lifestyle, accessories, and snack F&B brands tend to perform well due to the visual nature and impulsive buying behaviors on TikTok.
2. How long should a TikTok live session be?
30–60 minutes is the sweet spot for engagement. Ensure you're offering fresh content or deals throughout the stream.
3. Do I need a large following to start TikTok Live Selling?
Not necessarily. TikTok’s discovery engine helps surface your live to the right audience. Paid media can boost this further.
4. How do I measure success from my TikTok livestreams?
Track watch time, viewer retention, clicks, conversions, and comments. Then retarget those who engaged using TikTok Ads.
MajorForm helps brands connect emerging placements with their wider performance marketing strategy. If you are considering ChatGPT Ads in Singapore or across Southeast Asia, talk to our team about a focused test: what to advertise, how to measure it and what would justify scaling.
Contact us here.
