Reading time: 14 minutes
SEO for banks and fintech now covers what ChatGPT and Google say about accounts, cards, credit and insurance. On these topics, however, AI demands more reliable sources and still gets rates wrong. So the work runs through clear product pages, current data and the advertising rules of US and UK regulators.
In short
- On money questions, Google says AI Overviews demand even more reliable sources and, in many cases, point people to an expert.
- In a Yext study of local searches, brand websites made up 48.2% of AI citations in finance.
- AI gets rates wrong: in a Which? test, ChatGPT gave a British bank a savings rate it did not pay.
- In the US, FINRA, the CFPB, the FDIC and state insurance law reach the brand’s own pages, which AI reads. In the UK, the FCA does.
Does AI recommend banks, fintech and insurers?
Yes, mostly when people ask for a list, such as “best high-yield savings account.” In these cases, the answer usually names brands and sources. On money topics, however, AI companies say they act with more care. Google demands more reliable sources, and OpenAI says ChatGPT is no substitute for a licensed financial adviser.
Google treats these questions as YMYL (Your Money or Your Life). According to Google’s document on AI Overviews and AI Mode, they have “an even higher bar” for showing information from reliable sources. In addition, for many of these questions, the overview says when to seek an expert or check the facts.
OpenAI takes a similar line. It responded to a test by Which?, a UK consumer group. The company said ChatGPT search gives links so users can check the sources. Likewise, it said ChatGPT is no substitute for a licensed financial adviser. In other words, AI answers, but it tends to point to sources and urge caution.
Also, there is no broad paid shortcut. In an archived copy of OpenAI’s ad policy from July 2026, ads for financial products and services were restricted, and OpenAI said it might allow approved financial advertisers. The policy changes often, so check the current version. The details are in our guide on whether ChatGPT ads are worth it. Therefore, SEO for banks and fintech stays, first of all, organic. If the concept is new to you, start with our guide on what GEO is.

Why does financial content have its own rules (YMYL)?
Because a mistake about money can hurt the reader’s finances. That is why Google calls these topics YMYL and gives more weight to trustworthy content on them. In the US and the UK, moreover, regulators set advertising rules for financial products.
According to Google Search Central, its systems give extra weight to content with strong E-E-A-T on topics that could significantly affect people’s financial stability. In practice, SEO for banks and fintech needs a named author, sources and data that readers can check. Our E-E-A-T guide explains how to show experience and trust in your content.
This article, however, is not legal or financial advice, and it does not recommend any financial product or investment. NotHuman sells AI-visibility tracking, so we have a commercial interest in this topic. The rules below come from the official texts, while the GEO takeaways, marked as our reading, are ours. So, before you change any page, take your questions to legal and compliance. To compare with health, law and retail, see what changes in AI answers in each industry. Law has its own rules too; see the guide to GEO for lawyers.
Which rules apply to bank SEO pages in the US and UK?
The table lists the main ones. Each applies to the firm’s own pages and ads, which is what AI reads. The sections below explain them.
| Rule | Applies to | What it requires |
|---|---|---|
| FINRA Rule 2210 | Broker-dealers | Fair and balanced communications |
| CFPB UDAAP standard | Consumer finance providers | No deceptive claims or omissions |
| Regulation Z, 12 CFR 1026.24 | Closed-end credit ads | The APR when terms are stated |
| Regulation DD, 12 CFR 1030.8 | Deposit accounts | The APY and its conditions |
| FDIC, 12 CFR Part 328 | Any person, fintech included | No false deposit insurance claims |
| NAIC Model 880 | Insurers (state law template) | No untrue or misleading ads |
| FCA financial promotions | UK firms | Clear, fair, not misleading |
Credit and cards in SEO for banks: APR and rates
Under the CFPB’s Regulation Z, an ad for closed-end credit may state only terms the lender actually offers. If the ad states a finance charge rate, it must call it an “annual percentage rate” (APR). Also, stating the payment amount, the number of payments or the finance charge triggers more disclosures. These include the repayment terms and the APR.
For cards, section 1026.16 of the same regulation works in a similar way. According to the CFPB’s commentary, even a claim like “no annual membership fee” can trigger more disclosures. In our reading, therefore, a page that shows only a teaser rate gives AI incomplete data to repeat.
The CFPB also applies a broader test. Its UDAAP procedures say a provider of consumer financial products may not use deceptive practices. An act is deceptive when it misleads or is likely to mislead, the consumer’s reading is reasonable and the point is material. The CFPB presumes that information on costs, benefits and restrictions is material. For this reason, omitting a material limitation from an offer can be deceptive. In our reading, absolute claims like “guaranteed approval” also deserve a compliance check, because AI may repeat them.
Accounts and investments: what SEO for banks must show
For deposit accounts, Regulation DD says an ad cannot be misleading or inaccurate. If it states a rate of return, it must use the term “annual percentage yield” (APY). Where they apply, it must also state the conditions. Examples are the minimum balance, how long the APY is offered and, for variable-rate accounts, that the rate may change. Thus, in SEO for banks and fintech, each account page shows the APY, its conditions and the update date as text.
Fintech brands face another rule. Under the FDIC’s rules, no person may represent or imply that an uninsured product is FDIC-insured. A non-bank that talks about deposit insurance must also say clearly that it is not an FDIC-insured bank. When it implies that a product is insured, it must name the insured banks that hold the deposits. In our reading, AI may repeat “FDIC-insured” for an app that is not a bank, so the page should name the partner bank in text.
For investments, FINRA Rule 2210 says broker-dealer communications must be fair and balanced. They may not omit a material fact or make a false, exaggerated, promissory or misleading claim. They also generally may not predict or project performance. A registered principal must approve a retail communication before its first use. Retail communications include electronic ones made available to more than 25 retail investors in 30 days, so a public web page can qualify.
UK rules for bank SEO: the FCA
In the UK, the FCA says all financial promotions must be clear, fair and not misleading, whatever the media type. It names a website, a Facebook post and a tweet as examples. Specific rules then depend on the product. In our reading, a UK bank or fintech should treat every product page and social post as a promotion.
Insurance: what state law requires
In the US, insurance trade practices are defined state by state. The NAIC publishes a template for state laws, the Unfair Trade Practices Act (Model 880). It lists as unfair any statement or illustration that misrepresents the benefits, advantages, conditions or terms of a policy. It also covers any ad, including an internet posting, that is untrue, deceptive or misleading. States adopt their own versions, so check the states where you sell.
Across all these sectors, the principle is the same: what the brand publishes counts as the brand’s own claim. In our reading, then, SEO for banks and fintech, and for an insurer, starts with pages that show the right rate, APR or APY, coverage and terms.
Which sources does AI use for banks: comparison sites, regulators or forums?
No AI company publishes the list of sources it uses in finance. Still, the available studies show that the brand’s site matters in local searches. They also show that, on general questions, AI cites forums and videos too. The finance-citation rankings we found are vendor-run and use different prompt sets, so we do not repeat them.
Yext analyzed 6.8 million citations from ChatGPT, Gemini and Perplexity in location-based questions. As a result, in finance, 48.2% of citations came from brand websites, such as local bank branch pages. Yext, however, sells listings management. It also collected the data at locations worldwide and did not split the sample by country.
On general questions, though, the picture changes. In July 2026, Which? published a test of five AI tools with 15 personal finance questions in the UK. In it, Reddit and YouTube showed up as sources in several answers, and AI Overviews even cited an Instagram reel. Also, on travel insurance questions that did not say the user was in the UK, some answers gave US-centric information. Google, however, replied that 15 questions are not a representative sample.
That is why, in SEO for banks and fintech, your own page needs to answer the whole question, with local context. To map the sites AI uses in your market, see the sources AI cites most.
What the CFPB publishes about your brand
The CFPB, for example, runs a public Consumer Complaint Database. It publishes a complaint after the company responds, or after 15 days, whichever comes first. Anyone can download the data. The CFPB does not publish complaints it refers to other regulators, such as those about banks with less than $10 billion in assets.
We found no documentation that says whether AI uses this database. Even so, it is public data about your brand, and SEO for banks and fintech needs to know what it shows. The same goes for review sites and comparison sites, so check what they say before you measure AI.
How do you fix a wrong rate in an AI answer with SEO for banks?
Find the source of the error, fix that source and log the case. After all, AI usually repeats a page with an old rate, from the brand or from third parties. And we found no direct channel at AI companies for a brand to correct an answer.
In fact, this has already happened with real banks. In March 2026, a Which? savings expert asked the free version of ChatGPT for the best savings rate and where to put £25,000. It is a UK case, but the mechanism applies anywhere. In its answer, the AI said Marcus by Goldman Sachs paid 4.5% a year, when it paid 3.75%. It also left out that another bank’s rate fell to 2.25% after a bonus.
According to Which?, the problem was ChatGPT’s habit of using sites with rates that may be out of date. Moreover, the same question, asked on other occasions, brought different answers and sources.
What are the steps of SEO for banks to fix a wrong rate?
Therefore, in SEO for banks and fintech, the plan has six steps:
- Take inventory. List products, rates, APR or APY, fees and coverage, each with its effective date.
- Make the product page the single source. Then put the rate with the APR or APY, the conditions and the update date in text.
- Clean up old offers. Update or remove pages, PDFs and posts with expired rates.
- Ask third parties to correct. After all, comparison sites, partners and the press also publish the rate.
- Log the case. Save the prompt, the date, the answer and the cited source, because compliance will ask.
- Ask again. Finally, after the fix, run the same prompt again on more than one engine.
To find the page behind the error, follow our guide on when ChatGPT makes up information about your company. In healthcare, the same routine applies, as the guide to GEO for clinics and doctors shows.
Can a wrong rate in AI become a regulatory problem?
In our reading, the bigger risk in SEO for banks and fintech lies in the brand’s page, not in the answer itself. After all, the rules above cover the firm’s own offers and ads. We found no rule in them about third-party AI answers. However, if the wrong rate comes from an old brand page, that page is still published. Under Regulation Z, an ad may state only terms the lender actually offers. So check each case with legal and compliance.
How do you track SEO for banks by product?
Build a fixed list of questions by product, audience and intent. Then run each one several times on more than one engine. Note who shows up, which source was cited and whether the rate is right. Finally, keep presence apart from accuracy, because showing up with the wrong rate is also a problem.
| Product | Sample prompt |
|---|---|
| Account | “best high-yield savings account for an emergency fund” |
| Card | “credit card with no annual fee for fair credit” |
| Credit | “how to compare the APR of personal loans” |
| Insurance | “car insurance for a used car in Ohio” |
| Investment | “how to open a brokerage account as a beginner” |
| Fintech | “is a neobank account FDIC insured?” |
| Brand | “what is the annual fee on Bank X’s card?” |
For each answer, then, check the rate, APR or APY, fee, coverage and limits against the official page. Name the country or state in the prompt. In the Which? test, some answers to questions that did not say the user was in the UK gave US-centric information. Brands that sell in more than one market should test each one. Also, include criteria questions, like the APR one, because that is where AI explains the product.
Meanwhile, how many runs to do and how to build the report are in our guide on how to measure brand visibility in AI. In SEO for banks and fintech, the difference lies in the product matrix and in checking regulated data.
Frequently asked questions
A fintech follows many of the same advertising rules as a bank, but which ones depend on its products and license. Credit ads follow Regulation Z, and the CFPB’s UDAAP standard applies to the consumer finance providers it covers. A fintech that is not a bank also follows the FDIC’s rules on deposit insurance claims. It must not imply that it is insured, and it should name the partner bank. Its charter and license decide the rest, so ask legal.
A bank or an insurer can advertise on ChatGPT only in limited cases. In an archived copy of OpenAI’s ad policy from July 2026, ads for financial products and services were restricted, with approved financial advertisers possibly allowed. Check the current policy before you plan a budget. So, for most financial brands, the AI route is SEO for banks and fintech, which is organic.
AI names banks and products when the question asks for options, but AI companies urge caution. Google says that, for many money questions, the AI Overview says when to seek an expert. OpenAI, for its part, says ChatGPT is no substitute for a licensed financial adviser.
Main sources
- FINRA, Rule 2210, communications with the public.
- Consumer Financial Protection Bureau, unfair, deceptive or abusive acts or practices (UDAAP) examination procedures (2023).
- FDIC, 12 CFR Part 328, Subpart B, false advertising and misrepresentation of insured status.
- NAIC, Unfair Trade Practices Act, Model 880 (2024).
- Financial Conduct Authority, financial promotions and adverts (UK).
- Google, AI Overviews and AI Mode in Search (2025).
- Which?, test of five AI tools with 15 personal finance questions (2026).
- Yext, study of 6.8 million AI citations (2025).



