Can insurance companies advertise on ChatGPT today?
Insurance companies can advertise on ChatGPT in the US if OpenAI approves them case by case. Outside the US, OpenAI's ad policy calls financial services ads "generally prohibited".
That is OpenAI's ad policies page as we read 26 October 2026 (v1.6, September 2026), and it has changed before. This guide is for marketing and compliance leads at insurers and banks in Singapore, Malaysia, Australia, Canada and the US: sort each product line into "apply now" or "build readiness", then score your team.
So a Singapore insurer can build a campaign in Ads Manager and still be stopped at review. Our May analysis of the self-serve launch said financial services were excluded and Singapore was not live. Both have moved, as the next sections show.
Our position: waiting for stable rules is a plan to be late. Build the compliance sign-off path now with a ChatGPT Ads partner. Most Singapore and Malaysian groups have no US entity; for them this quarter means drafting, not spending.
What does OpenAI's ad policy say about financial services?
OpenAI's ad policy lets approved US advertisers in ten financial categories apply case by case, and calls such ads "generally prohibited" outside the US.
Advertisers "may be required to provide proof of licensure." The ten categories are:
- Insurance
- Deposit accounts
- Credit cards
- Personal loans
- Payment services
- Mortgages
- Auto loans and leasing
- Investment services and brokerages
- Financial planning
- Credit monitoring
A few clauses matter more than they look:
- Banned everywhere: credit repair, debt settlement, debt assistance, and alternative investments such as bullion.
- Health insurance follows the same US-only rule, under health services.
- Tools and education that promote no financial product are "generally permitted". But advertisers whose "primary business model or offerings fall within restricted or prohibited categories" may be ineligible. The policy's examples are healthcare and weapons, so how the clause treats an insurer is unclear. Do not attempt an education campaign without OpenAI's approval in writing.
- Two layers of text. The general content paragraph still says financial services were "disallowed at launch". The specific financial services section lower down carries the current rule.
Placement is a separate question from eligibility.
Since April 2026, medical, legal and financial advice conversations are no longer blocked from ads by default, per the changelog and trade coverage. Sensitive contexts stay off limits: OpenAI's policy names emotionally reliant conversations and mental and personal health chats, and its Help Center says ads are not eligible near personal health, mental health or politics. Do not plan on a term-life explainer carrying your product ad. OpenAI has not said that, and in June a spokesperson told Marketing Brew the company was not widely placing ads in financial conversations, naming budgeting and financial-education tools as the kind of ad that can appear.
Which financial products are eligible in which markets?
Only the US allows insurance and banking ads, case by case; OpenAI's policy page rules them out elsewhere, including Singapore and Malaysia.
| Category | United States | Singapore, Malaysia, Australia, Canada | Page stamp |
|---|---|---|---|
| Insurance | Case by case Licensure proof may be required | Generally prohibited | 10 Sep 2026 |
| Health insurance (listed under health services) | Case by case | Generally prohibited | 10 Sep 2026 |
| Deposits, cards, personal loans, payments, mortgages, auto loans and leasing, investments and brokerage, financial planning, credit monitoring | Case by case | Generally prohibited | 10 Sep 2026 |
| Tools and education promoting no financial product | Generally permitted Primary-business clause may apply | Generally permitted Same caveat | 10 Sep 2026 |
| Credit repair, debt settlement, debt assistance, bullion | Not allowed | Not allowed | 10 Sep 2026 |
"Case by case" is a queue, and the policy's review section says restricted categories may face enhanced advertiser verification or manual review. "Generally" leaves room for exceptions. It is still no invitation to test the boundary with your brand on the account. So the honest answer to "can insurance companies advertise on ChatGPT" depends on the row. Outside the US the policy says "generally prohibited", not "prohibited", so ask OpenAI in writing before you assume a flat no.
A regional insurer sells travel and term life cover in Singapore and Malaysia, with no US entity. That is four product-market rows, and all four are closed today. There is no media decision to make this quarter.
There is a timing decision.
Say its compliance committee needs eight weeks to approve any new ad format. If the policy opens Singapore with no notice, an insurer starting that day is live eight weeks later at best. One that drafted its standard now, mostly by adapting its search ad rules, waits only for OpenAI's own review.
Not sure which of your product lines sit in an open category? Send us your product list and the markets you sell in through the contact page, and Leapbuzz returns an eligibility map for Singapore, Malaysia, Australia, Canada and the US.
This post describes platform policy as published. It is not legal or regulatory advice. Local promotion rules can apply on top of any platform approval; our fintech marketing compliance guide for APAC covers them by market, and your compliance team makes the call.
What did the 23 September 2026 Southeast Asia rollout change?
On 23 September 2026 OpenAI announced that ChatGPT Ads would begin rolling out in Indonesia, Malaysia, the Philippines, Singapore, Thailand, Vietnam and Taiwan. OpenAI's Ads Manager availability page, read 26 October 2026, lists Singapore and Malaysia as Available. The finance category did not open with them. On 21 October 2026 OpenAI said in its "ChatGPT Ads expands to 26 new countries" post that ChatGPT Ads had reached 89 countries, 26 more (read 26 October 2026); none of the reporting we found says the finance rule changed, and the policy page as read on 26 October still carries the outside-the-US line.
Four details shape the rollout:
- Regional coverage (Marketing-Interactive) dated the start to 24 September local time.
- Trade press reported self-serve open in all seven from day one; user delivery is a rollout, not a finished market.
- Ads show only to Free and Go plan users, per the 21 October post.
- OpenAI's Ads Manager availability page also requires the advertising legal entity to be based in a listed country.
The less obvious read: the rollout raises the cost of waiting for a Singapore insurer, even though it cannot advertise yet. Other categories can buy through self-serve as delivery rolls out, so if the financial ban lifts later, those advertisers will already know the format. A pre-cleared creative and landing-page file shortens your internal review. It does not shorten OpenAI's, and it does not mean the category will open.
This is our reading of platform policy, not legal or regulatory advice. A regional group might think of running Singapore campaigns through its US subsidiary. OpenAI's advertiser policy requires compliance "in the regions where they operate and target advertising" and bars misrepresenting "business location, service area, or eligibility to advertise in particular markets." Ask your compliance team how local promotion rules treat an ad bought through a foreign entity before anyone tries it.
How often does OpenAI change its ad policy?
OpenAI's ad policy changelog ran from v1.0 in March 2026 to v1.6 in September 2026: seven versions, v1.0 to v1.6, four of them touching finance or consumer protection.
The page moved often; the finance rule itself changed in April and July. Anyone who answers "can insurance companies advertise on ChatGPT" from a spring briefing is working from an old rule. The chart counts every version; the table lists the finance entries.
| When | Change | Where it is reported |
|---|---|---|
| March 2026 (v1.0) | Ad policies first published; financial services "disallowed at launch" | OpenAI ad policies page changelog (read 26 October 2026) |
| April 2026 | Medical, legal and financial advice conversations no longer blocked from ads by default | OpenAI changelog; Search Engine Journal, 12 Aug 2026; Marketing Brew, 17 Jun 2026 |
| July 2026 (v1.3) | Financial and health sections updated "to clarify eligible categories and markets" | OpenAI changelog; Search Engine Journal dated the page update to 10 August |
| September 2026 (v1.6) | Right to decline ads that conflict with OpenAI's principles, business interests or competitive position | OpenAI changelog |
The signal came earlier than August. Marketing Brew reported the case-by-case language and a widening financial-services pilot on 17 June 2026; the July entry formalised the category list, and Search Engine Journal wrote it up on 12 August. The September entry matters too: approval in a listed category is still discretionary.
Why build the sign-off path before you are eligible?
Build it now because your own compliance review, not OpenAI's queue, is usually the slower step, and the policy page was revised six times between March and September 2026.
When a category opens, OpenAI's review is one queue and your compliance committee is another.
A conversational placement raises questions your search and social templates never answered:
- What may a short ad say next to a question about claims?
- Who signs off when the context is a live chat nobody at the bank will ever see?
Those answers take committee time, and they cost no media budget to settle now.
A CMO with this quarter's targets will object to spending committee time on a channel that bars the category. Fair. The work is smaller than it sounds: mostly your existing search advertising rules, adapted for a short ad beside an AI answer, plus a licence file.
For most new ad surfaces we would tell a bank to wait and let someone else make the first mistake. Here we disagree, with one exception: if you lack both a creative standard for conversational ads and a way to measure incremental results, do not chase inventory. Build those two first. This is a bet on timing, not a forecast: if the rule stays US-only for a year, the cost is committee hours.
"We're expanding the pilot deliberately, with a focus on learning and getting the experience right for users."
Asad Awan, head of monetization, OpenAI, told Marketing Brew on 20 May 2026
The organic side is open everywhere. ChatGPT answers insurance and banking questions in Singapore and Malaysia today, ads or no ads, and OpenAI's stated plans for the ad platform point to more markets. Whether your brand appears in those answers is work you can start now, and our ChatGPT readiness work pairs it with ad preparation for restricted sectors.
Is your bank or insurer ready? Score the seven steps
Apply this quarter if you have a licensed US entity in a listed category and at least five of the seven steps below done. Otherwise, finish the steps. Leapbuzz, an AI performance marketing consultancy in Singapore working across five markets, built these steps from OpenAI's published policy for its ChatGPT Ads readiness service.
Nothing ticked yet: 0 of 7. Without a US entity in a listed category, your job this quarter is readiness, not spend.
Rule: with the US box ticked and 5 or more steps done, apply and plan a small capped test. With the US box ticked and fewer than 5, finish the missing steps first. Without it, keep building and re-check the policy monthly.
Insurers weighing a deeper presence than ads should also read our take on building an insurance app in ChatGPT. Our insurance and banking and finance pages cover the wider channel plan.
Start with step 1 at the next month-start. If that read shows your category open in a market you sell in, steps 3 and 5 decide whether you are live that month or next quarter.