On a Sunday evening, the digital head of a private bank asks ChatGPT on her phone how a family should save for a child’s college fees. The answer runs to six paragraphs and names two mutual fund categories and a recurring deposit. Below it sits a small card marked sponsored, from a budgeting app. Her bank’s education loan appears nowhere on the screen. On Monday her chief marketing officer wants to know two things about ads in ChatGPT: can the bank buy that card, and should it?
What are ads in ChatGPT and other AI assistants?
Ads in ChatGPT and other AI assistants are paid messages that a chatbot shows next to its own answer, marked as sponsored and kept visually apart from the reply. The platform picks the ad from the subject of the conversation, so a question about saving for school can trigger an ad for a savings tool. The advertiser pays for the placement and has no say over the answer itself.
The category is new. OpenAI began testing ads on the Free and Go tiers of ChatGPT on 9 February 2026, according to NBC News, and said the ads would be clearly marked, kept separate from answers and shown only to adults. TechCrunch reported on 27 August 2026 that OpenAI would start showing ads to Free and Go users in India, beginning with 50 brands and with WPP and Omnicom as agency partners.
Three other names come up in the same conversation. Google shows ads above, below and inside its AI Overviews in some markets, and has added a sponsored listing to AI Mode. Microsoft runs ad formats inside its Copilot assistant. Perplexity tested sponsored follow-up questions from November 2024 and has since stepped back from advertising.
The working definition for a financial marketer is short. An AI assistant ad is a labelled paid unit triggered by conversational context, sitting on a page whose main content the advertiser does not control.
How do sponsored AI answers differ from search ads and from GEO?
A sponsored AI answer differs from a search ad in what triggers it and what surrounds it, and it differs from generative engine optimisation (GEO) because you pay for it. Those two differences decide where your compliance effort goes.
A classic search ad is bought against keywords. You choose the phrase, write the headline and the description, and the ad appears in a results list among other links. The page around it is a list your brand can also rank in. In an AI assistant, the trigger is the whole conversation. Microsoft said in October 2024 that Copilot ads would be triggered by the whole conversation within a single session, not only the last prompt. The page around the ad is a written answer that names products, explains trade-offs and may cite sources.
That written answer is the part GEO works on. GEO, sometimes called answer engine optimisation (AEO), is the practice of making your own pages clear and citable so that AI engines quote them correctly in the free answer. Our AI Search and GEO practice covers this side. No money changes hands with the engine. You earn the mention by publishing facts an engine can read and trust.
The practical difference shows up in three places:
- Who writes the words. In a search ad, you write every word. In Google’s AI Mode, MediaPost reported in May 2026 that the copy next to a sponsored listing appeared to be written by AI, not the advertiser. A regulated brand has to plan for copy it did not draft.
- What sits next to you. A search ad sits next to blue links. An AI ad sits next to an answer that may name your rival, quote an old interest rate for your own product, or recommend a category you do not sell.
- What you can target. Search ads let you pick keywords and exclude others. Google says advertisers cannot directly target or opt out of AI Overview placements. You get context matching and lose some control.
- What you can measure. Google says it does not yet offer separate reporting for ads shown inside AI Overviews. Plan your test with the reporting the platform actually gives you, which today is thinner than search.
The verdict for a bank or insurer: treat ads in ChatGPT and its rivals as a context buy with search-style copy rules, and treat GEO as the work that decides what the answer above your ad says.
How do ads in AI assistants work under Indian and Gulf rules?
Under Indian and Gulf rules, treat an ad in ChatGPT or any other AI assistant as an advertisement like any other, and plan for every disclosure your regulator requires on a banner or a search ad. We have not found a rule from any regulator in either region written specifically for chatbot ads. The existing codes are media-neutral, and the platform adds its own gate on top.
SIGNALNewsletter by Yamm Labs
The next piece, straight to your inbox.
New writing on BFSI brands, B2B and influencer marketing, and AI search.
One click to unsubscribe. Privacy policy
The platform gate comes first. According to Marketing Brew’s review of OpenAI’s ad policy in 2026, financial services ads are allowed only from approved advertisers, approved manually and case by case, and sensitive conversations stay barred from ads. Google says it does not show ads within AI Overviews for sensitive verticals, and it names finance among them. So a bank’s first question is not about copy at all. It is whether the platform will accept a financial advertiser from India in the product category it wants to promote. Check the platform’s current eligibility list for your country before anyone writes a line.
If the platform says yes, the Indian regulator’s rules decide what the ad and its landing page must carry. Here is how that works by sector, described at the level we can confirm:
- Mutual funds. The Securities and Exchange Board of India (SEBI) requires the standard mutual fund risk warning in advertisements, and its 2008 circular on the standard warning says that in audio-visual ads the warning must be displayed and voiced for five seconds. If an assistant ever reads sponsored content aloud in voice mode, that rule is the one to test against first.
- Insurance. The Insurance Regulatory and Development Authority of India (IRDAI) replaced its 2021 advertisement regulations with its 2024 regulations on protecting policyholders’ interests. Every insurer’s ad must show its registered name with its trade name or logo prominently, and an ad issued by a distribution channel needs the insurer’s written approval before it goes out. Ads are no longer filed with IRDAI, so the insurer’s own approval record is the evidence.
- Loans. The Reserve Bank of India (RBI) requires lenders to give a Key Facts Statement (KFS), including the annual percentage rate (APR), before a retail or MSME term loan contract is signed. Those rules are now consolidated in RBI’s Master Directions. The ad itself is short, so the KFS logic lands on the journey after the click.
- Everyone. The Advertising Standards Council of India (ASCI) code asks that ads inviting the public to invest must not mislead about returns or security, and that claims of fact can be substantiated. Its rules for ads on media handles ask that they be clearly distinguishable from editorial content, with a label such as Sponsored shown upfront.
The Gulf works the same way at the principle level. In the UAE, the Central Bank of the UAE (CBUAE), the Securities and Commodities Authority (SCA) and, inside the Dubai International Financial Centre, the Dubai Financial Services Authority (DFSA) each set their own consumer protection or financial promotion rules. In Saudi Arabia, the Saudi Central Bank (SAMA) and the Capital Market Authority (CMA) do the same. Which one applies depends on your licence. We describe these generically on purpose, because the exact wording differs by regulator and your compliance team should read the current text for the licence you hold.
For a global contrast, the UK Financial Conduct Authority (FCA) says on its financial promotions page that all promotions must be clear, fair and not misleading regardless of the media type. That phrase is the cleanest summary of how most regulators will read a chatbot ad. The format is new and the test is old.
This is general guidance, not legal advice. Your compliance team decides what each ad must carry.
What regulator-driven decisions does an AI ad force on the page?
An AI ad forces decisions on the landing page and the product page, because the ad unit is too small to hold the disclosures and the answer above it is built from your pages. These are the decisions we would put in front of a compliance reviewer before any spend.
- Put the full disclaimer on the first screen after the click. The card holds a headline and a link. Put the SEBI warning, the insurer’s registered name with trade name or logo, or the lender’s APR on the first screen of the landing page on a phone, where the reader sees it without a scroll or a tap to expand.
- Name the legal entity in plain text at the top of every product page. An engine that cannot tell your registered name from your brand name may describe your product under the wrong entity. Write the registered name, the brand name and the regulator that supervises you in one sentence near the top.
- Keep every regulated fact in one dated table. Rates, fees, charges and the risk line go in a plain HTML table with an “as of” date and a named owner who updates it. Engines quote what they can read. A rate inside a PDF brochure or an image is the one they get wrong.
- Mark the product page up with the right schema type. Schema.org has types for a FinancialProduct, a LoanOrCredit and an FAQPage. Markup does not guarantee a mention, but it removes guesswork about what the page describes.
- Record which source each claim cites. Keep a short claims register: the claim, the source document, the approver and the date. When an engine quotes your page, you can prove where the figure came from.
The order is deliberate. Steps two to five make the free answer accurate, and step one makes the paid click safe. If you only have budget for one, do the free side first, since it is the part every user sees whether or not your ad runs.
Where do financial marketers meet ads in AI assistants in practice?
Financial marketers meet AI assistant ads in four places today: ChatGPT, Google’s AI Overviews and AI Mode, Microsoft Copilot, and the trade press covering Perplexity’s exit. Each sets a different rule for finance, so the plan differs by platform.
ChatGPT. Ads appear on the Free and Go tiers and are labelled and set apart from the answer, according to NBC News. In India, TechCrunch reported in August 2026 that an ads manager would follow the next month, with a minimum daily budget of ₹725 per campaign. OpenAI also said in February 2026 that it had more than 100 million weekly ChatGPT users in India, a company-reported figure. For ads in ChatGPT, financial advertisers face the case-by-case approval described above, and OpenAI says it keeps conversations private from advertisers.
Take a hypothetical test at that floor. A daily minimum of ₹725 over a 30-day month is ₹21,750. That is a small line in most bank media plans. The non-obvious lesson is that money is not the constraint. The constraint is whether your category is approved and whether your landing page passes your own compliance review in time.
Google AI Overviews and AI Mode. Google’s help page says ads can show above or below AI Overviews in every market where Overviews run, and inside the Overview in English in a short list of countries that includes India. It also says finance is one of the sensitive verticals where ads do not show within the Overview. In AI Mode, MediaPost reported in May 2026 a sponsored listing placed in the middle of organic recommendations, carrying a Sponsored tag. For a bank, the plain reading is that the inside-the-answer slot on Google is closed to finance for now, and the slots above and below work much like search.
Microsoft Copilot. Microsoft’s advertising blog said in October 2024 that Copilot ads would be matched to the whole conversation and introduced by a short “ad voice” line that links the ad to the answer. That line is written by the platform. A compliance reviewer should see examples of it before approving a campaign, because it is the sentence closest to your brand that you did not write.
Perplexity. TechCrunch reported in November 2024 that Perplexity began testing ads as sponsored follow-up questions in the United States, labelled as sponsored, with answers still generated by Perplexity. MacRumors, citing the Financial Times, reported in February 2026 that Perplexity was stepping away from advertising over concerns about user trust. Treat it as an engine to optimise for, not one to buy.
In the Gulf, check each platform’s own availability list for the UAE and Saudi Arabia before planning. These products have launched market by market during 2026, so a third-party country count can be out of date within weeks.
The ad buys a place under the answer; only your own pages decide what the answer says about you.
Should a financial brand buy AI assistant ads or invest in GEO first?
A financial brand should invest in GEO first and test ads second, because finance is restricted on the ad side and fully open on the organic side. Every user sees the free answer. Only some see the ad, and only if your category clears the platform’s review.
Choose to test ads now if your product is in a category the platform clearly accepts, such as a budgeting or financial education tool, if your landing page already carries the full disclaimer on the first screen, and if you can live with thin reporting for a quarter. Choose to wait if you sell credit, insurance or investment products that need case-by-case approval, or if your product pages still hold rates inside PDFs.
A short test plan for ads in ChatGPT or a similar assistant:
- Confirm eligibility for your country and category in the platform’s own policy pages, and save a dated screenshot.
- Get one landing page approved with the full disclosure on the first screen.
- Run at the platform minimum for 30 days with a single message.
- Measure what you can: impressions, clicks and approved applications from that landing page.
- Stop if the platform’s own copy, such as an AI-written explainer, says anything your compliance team would not have approved.
For the second group, the work is on your own site. Our GEO and AEO FAQ for financial brands answers the common questions on what engines read, and the AI visibility service for BFSI brands describes how we audit and fix it.
What to tell your compliance team, in a form you can paste into an email: “We propose a 30-day test of a sponsored placement in an AI assistant. The ad states only the product name and one approved line. The landing page carries the full regulatory disclaimer on the first screen. The platform may add its own explainer text, and we will review samples before launch and stop the test if that text makes a claim we have not approved. The approval record sits in our claims register.” This is general guidance, not legal advice; your compliance team decides.
Related terms a financial marketer should know
These are the terms that come up when a team plans for ads in ChatGPT and other AI answers. Each is defined in one line with where you meet it and why it matters for a regulated brand.
| Term | Plain meaning | Where you meet it | Why it matters |
|---|---|---|---|
| Sponsored placement | A paid, labelled unit shown beside or below an AI answer | ChatGPT Free and Go, Google AI Mode, Copilot | It is an advertisement in law, so every disclosure rule applies |
| Organic answer | The unpaid text the assistant writes in reply | Every AI assistant | It is built from public pages, including yours, and users read it first |
| Generative engine optimisation (GEO) | Making your pages clear and citable so engines quote them | Your website and product pages | It decides whether the free answer names you, and with which facts |
| Answer engine optimisation (AEO) | A close cousin of GEO focused on direct-answer formats | FAQ pages, short answers, schema | Short, quotable answers are what engines lift whole |
| Contextual trigger | The platform picks an ad from the conversation topic, not a keyword | ChatGPT and Copilot ad setup | You lose keyword exclusions, so check where the ad can appear |
| Ad voice | A platform-written line that links the ad to the answer | Microsoft Copilot | It is copy near your brand that your team did not approve |
| AI Overview | Google’s AI summary at the top of some search results | Google Search | Google says finance ads do not show inside it |
| Citation | A link the engine shows as the source of a statement | ChatGPT search, AI Overviews, Perplexity | A cited page is the one whose rate or fee gets quoted |
| Key Facts Statement (KFS) | RBI’s standard summary of a loan’s cost, including APR | Loan journeys after the click | An AI ad for a loan leads to a journey that must deliver it |
| Standard warning | SEBI’s mandatory mutual fund risk line | Every mutual fund advertisement | Its audio-visual rule is the test case for any voice ad |
| Claims register | A log of each claim, its source, approver and date | Your compliance file | It proves where a quoted figure came from |
To see how these terms play out on a real site, our AI visibility log records what engines said about our own brand on day one, before any fixes.
People also ask
Can Indian banks and insurers advertise in ChatGPT today?
Not by default. Reports on OpenAI’s 2026 policy for ads in ChatGPT describe financial services ads as allowed only for approved advertisers, reviewed manually and case by case, in the US, and say financial ads outside the US are generally not allowed. Check the current eligibility list for India and your product category before you plan spend.
Do ads in ChatGPT change the answer ChatGPT gives?
OpenAI says ads do not influence ChatGPT’s answers, and the ads are labelled and kept separate from the reply. The answer is built from public sources, so the way to change what it says about you is to fix your own pages.
Does an AI assistant ad need the same disclaimer as a search ad?
Yes. Indian regulators and ASCI apply their rules to advertisements whatever the medium, and the FCA in the UK says the same about financial promotions. The ad card is small, so the full disclaimer has to sit on the first screen of the landing page.
Should a financial brand spend on AI assistant ads or on GEO first?
GEO first for most banks, insurers and fund houses. Finance is restricted on the paid side of most AI platforms, while every user sees the organic answer. Test ads once your category is approved and your landing page carries the full disclaimer.
Before you pay for a place under the answer, find out what the answer says about you.
Get a free Fintech Brand Audit across 8 dimensions, plus 3 fixes. No pitch. Run the free Brand Audit →