Formerly WebStrategies, Inc.

Christina Odom
Jul 14, 2026
Picture this: a potential member opens ChatGPT and types, "What's the best way to finance a used car?" ChatGPT gives its answer. And just below that answer, in a clearly labeled box, sits a sponsored ad. OpenAI began testing ads in ChatGPT for U.S. users in February 2026, and by May, it had opened an Ads Manager that lets businesses of nearly any size create campaigns directly.
If you hold the digital advertising budget at your credit union, you've probably already heard about this, and you're likely asking the obvious question: Should we be spending here?
Here's a clear-eyed look at how ChatGPT advertising actually works, what makes it particularly complicated for credit unions, and what we recommend doing instead as the platform matures.
If you've run traditional search ads through Google, set aside most of what you know. ChatGPT ads work differently in a few important ways.
There are no keywords. Instead of bidding on search terms, OpenAI matches ads to the topic of the conversation a user is having. If the user has ad personalization enabled in their settings, OpenAI may also draw on additional signals, such as their past chats, memory, and prior ad interactions. (That setting is currently on by default; users who turn it off see ads based strictly on the current conversation.) Either way, if someone is discussing auto loans, an auto lender's ad may appear. Advertisers are essentially betting on conversation topics, not keywords.
Ads appear below the answer, never inside it. Every ad is labeled as sponsored and visually separated from ChatGPT's response. OpenAI has been emphatic on this point: ads do not influence the answers ChatGPT gives. Advertisers cannot pay to change what the AI says.
Only some users see ads. Ads appear for people on the Free and Go (low-cost) plans. Paid subscribers (Plus, Pro, Business, Enterprise, and Education) see no ads at all. OpenAI also does not show ads to users under 18 or predicted to be under 18.
Privacy is walled off. Advertisers never see user conversations, chat history, or personal details. They receive only aggregated performance data, such as total views and clicks.
Buying is now open to almost everyone. The platform launched with a small group of large advertisers. As of May 2026, OpenAI offers a beta self-serve Ads Manager with both cost-per-impression and cost-per-click bidding, and the previous minimum spend requirement has been removed. In theory, any U.S. business can sign up. One note from our own testing: the platform does enforce a modest minimum daily budget - low enough that most advertisers will be comfortable with it, but enough that this isn't a channel you can dabble in with a token monthly spend.
However, for credit unions, there's a significant asterisk involved.
Technically, yes, but not easily. Financial services are a restricted category on the platform. Ads from financial institutions may be approved only on a case-by-case basis, through manual review, with enhanced advertiser verification. There is no guarantee of approval, and no published timeline for how long the review takes.
Even within the approved lane, some financial sub-categories are off the table entirely. OpenAI's current ad policies disallow ads for cryptocurrency, credit repair, debt settlement, and debt assistance services. That's worth pausing on for credit unions, specifically: debt counseling and credit-builder programs are core community offerings for many institutions, and under today's policies, those services can't be advertised on the platform even from an approved account.
Even after approval, delivery is limited. OpenAI has confirmed it is not widely showing ads in financial conversations. Ads may appear in those contexts only in limited circumstances where they make contextual sense. Translation: a credit union could clear the approval process and still find that its ads rarely serve, because the inventory itself is deliberately throttled.
And the ground is still shifting. In one telling example, OpenAI invited law firms to the platform in May 2026, then reversed course weeks later, excluding legal services entirely. Policies in regulated categories are being written and rewritten in real time. What's allowed today may not be allowed next quarter.
Not precisely. ChatGPT's targeting is built around conversation topics, with geographic targeting currently available at the DMA (media market) level. That's meaningfully better than national delivery, but it's a blunt instrument for credit unions. DMAs are built for television markets, and a single one can sprawl across dozens of counties. A community credit union whose field of membership covers three of those counties would still be paying for impressions across the entire market, reaching plenty of people who are legally ineligible to join.
Every dollar spent outside your field of membership is a dollar spent on a person who can never become a member. For a large national fintech, DMA targeting is workable. For most community credit unions, the gap between "media market" and "field of membership" means real waste until targeting controls tighten enough to close it.
The short answer: all of the ones you already follow. ChatGPT is a new medium, not a new rulebook; a new ad platform doesn't reset your compliance obligations, and every requirement your compliance team already enforces on Google or Meta follows you here.
A question we hear often: has the NCUA issued rules for AI advertising? In short, no, but it doesn't need to. NCUA's advertising rule (Part 740) defines an advertisement as a commercial message in any medium designed to attract public attention to a product or business. That definition clearly makes a ChatGPT ad an advertisement under Part 740.
That means the familiar requirements follow you onto the platform:
NCUA rules are only part of the picture. The rest of the familiar framework applies too:
ChatGPT may be a new medium, but the platform wasn't built with these requirements in mind. That gap between the rules and the ad format is exactly what needs to be mapped beforehand.
Not in the way your board will want to see. Right now, the platform's reporting is largely limited to upper-funnel metrics such as impressions, clicks, and engagement. From our own testing, that's the practical reality: you can see that people saw and clicked your ad, but connecting those clicks to funded loans or opened accounts inside the platform isn't there yet.
Conversion-optimized campaigns are on OpenAI's roadmap and beginning to roll out, but they're not yet widely available. HubSpot just launched a brand new integration for ChatGPT Ads, but it remains highly experimental and unvetted. For credit unions managing their marketing through HubSpot, this means ChatGPT ad activity won't seamlessly sync with your CRM the way your Google or Meta data does, at least not out of the gate. Measurement falls back on the basics: UTM-tagged landing pages and your own analytics, stitched together manually.
For a channel asking for real budget, "we can count clicks" isn't yet a business case. Until conversion measurement matures and connects to the systems credit unions actually report from, ROI on this platform is an estimate, not an answer.
Three things, none of which involve moving the budget to OpenAI just yet.
1. Keep your proven channels fully optimized. Your current high-performing channels are measurable, compliant, and targetable to your field of membership. Nothing about ChatGPT's launch changes that math today.
2. Invest in showing up in ChatGPT's answers, not its ads. Here's the part that matters most. Ads cannot buy their way into what ChatGPT actually says. When a potential member asks "What's a good credit union near me?" or "Should I refinance my auto loan?", the AI's organic answer is shaped by the content, structure, and credibility of what's published on the open web. Making your credit union citable in those answers through structured content, clear factual language, and technical optimization is work you can do right now. It's also the more durable asset: rented ad space disappears when spend stops; earned visibility in AI answers compounds. Visit the GEO for Credit Unions page for more information.
3. Let the testing happen before your budget does. Our ad team is actively vetting the platform in an internal environment, evaluating exactly how it handles traffic, targeting, compliance, and data privacy in practice. The platform is evolving constantly. Features that don't exist today may ship next month, and policies in place today may shift. We want to be fully confident in its performance and value before recommending any budget move. When the platform is truly ready, our clients will be positioned to move quickly and confidently, with the trial-and-error already done.
For credit unions right now, it's a platform where the approval process is manual and uncertain, ad delivery in financial conversations is deliberately limited, the targeting can't yet honor your field of membership, and every existing NCUA rule applies to an ad format that wasn't designed for them.
Your budget belongs on channels that are proven, compliant, and accountable, while the work of testing, mapping compliance, and building organic AI visibility happens in parallel. That's how your credit union ends up with the first-mover advantage without the first-mover risk.
Ready to prepare your credit union for the future of AI search and advertising? Successful AI ad strategies are built on a strong organic foundation. Before you invest in paid AI placements, lay the groundwork by understanding how AI engines perceive your brand today.

Let's build something measurable together.