Thinking Forward: What Happens When AI Becomes Your Members’ Financial Advisor?

In my role at Filene, I spend a lot of time thinking about how credit unions support their members’ financial health, so when my husband and I were looking for advice, we naturally turned to our credit union which offers both financial coaching and financial advisors.

We have never used either service. Part of it is scheduling: a form, a callback window, and weekday slots that assumed one of us could step away from work for a conversation we wanted to have together. The other part was fit. The coaching program is seemingly built for members working on the fundamentals; the advisor path was on the other end of the spectrum, asking about estates and managing wealth. Neither seemed to be for us, living somewhere in between.

A week later, on a Saturday at 9 p.m., we asked ChatGPT instead. No form, no weekday slot, and no perfect box to fit into to receive advice. In doing so, we joined a trend moving remarkably fast. Depending on the study, somewhere between 20% and 50% of Americans have asked an AI chatbot a personal finance question, covering advice on everything from savings and investing strategies to credit monitoring and debt management to retirement planning and insurance.¹ This is in addition, of course, to asking AI to recommend financial products, a topic explored in a previous Thinking Forward.

The instinctive response may be to educate members about why they shouldn’t be using AI for financial advice, but the reality is that it is happening and becoming more common. If you accept members using AI for financial advice as a reality, the task at hand becomes understanding how it might impact members’ understanding of their financial picture and preparing your staff for members who arrive already holding an answer, right or wrong.

Three ways financial advice from AI might impact your members

1. Members finally get help they were too embarrassed to ask for.

In our Financial Well-Being benchmarking survey, 22% of credit unions named awareness and stigma as a barrier to members getting support: members don’t know the service exists, feel embarrassed to ask, or fear they’re too far behind to recover. That shame keeps your members from getting help and acting on situations that get worse with time.

AI removes that shame, and the friction with it—since there’s no one on the other side, there’s no calculation of shame at all. A member can share how much debt they have, ask, and re-ask, all of the “stupid” questions, and explore different paths without wondering what someone is thinking about them.

Interestingly, if consumer behavior mirrors what happened when they got access to more information in the healthcare space, credit union expertise may become more in demand, not less. When WebMD went live, there was huge concern that the need for medical professionals would drop. Instead, people did have more access to information and did do more searching before speaking to medical professionals, but it didn’t decrease the demand; it changed the conversation patients were having with their doctors.

If banking follows healthcare, more-informed consumers will seek out experts to help them move forward. In this scenario, AI may turn out to be the most effective referral source your financial guidance program has ever had, reaching exactly the members who were too ashamed to seek help.

However, that referral only happens if the AI knows your program exists. This takes us back to April’s Thinking Forward, which explored what happens when AI agents choose financial products on a member’s behalf. The blog’s first recommendation, AI legibility, applies to your financial well-being offerings as much as to rate sheets.² Your well-being products and services need to be structured and specific enough that a model can surface them. A coaching program described in vague mission language, buried three clicks into a member services page, is unlikely to be surfaced by AI.

2. Members act on confident advice that doesn’t know their situation.

A September 2026 PensionBee study found 57% of Americans who ask chatbots for money advice would act on it without verifying, including on decisions that are hard to reverse.³ Nearly a quarter said a chatbot had already given them wrong information about their money.

One core problem is a lack of context. A general-purpose AI chatbot starts from a blank page. It knows only what the member shares, and people, even in the privacy of an AI chatbot, are unlikely to share complete or accurate pictures of their finances.

This, however, may change as chatbots such as ChatGPT are beginning to connect directly to financial accounts, which would close the context gap without a financial institution being involved. But connection requires permission and relies on trust that isn’t there yet. MX found that only 32% of consumers trust AI to help manage their finances, while 50% actively distrust it.^6,7 Among those asking for AI financial advice, PensionBee found 53% report moderate or extreme privacy concerns.⁴

The obvious answer is for credit unions to use their own AI chatbots, not to compete for the conversation, but because an assistant that can see the actual balances gives materially safer advice than one that can’t. However, two big questions emerge that need further exploration:

• Will members be as candid with an assistant that belongs to their financial institution? Part of why people speak freely to ChatGPT is that it has no stake in the outcome.

• How much better is the advice from a credit union’s chatbot if a member has their finances spread across multiple FIs? This is where ChatGPT could hold an advantage—in addition to connecting their credit union account, your members can connect every FI they’re working with, giving it a much more complete picture to work from.

3. Members know more and do exactly the same things.

Returning to the healthcare scenario, WebMD helped patients be better informed, but it didn’t change the central behavioral problem. Adherence to long-term treatment still averages roughly 50% in developed countries, the same place the WHO put it in 2003.⁵ More information did not lead to better outcomes.

We do not need to wonder if something similar will happen in the financial services space. There is already substantial evidence that a knowledge gap is not the main driver of poor financial health. The CFPB notably shifted their financial health framework away from literacy for this exact reason. One study that modeled financial well-being against financial literacy, financial socialization, self-control, and financial technology found that financial literacy on its own showed no significant direct effect on financial well-being.⁶ Better tools and better information only mattered when they changed what people did. Consumers receiving more financial knowledge through AI is unlikely to have a meaningful impact on their financial health.

The use of AI for financial information may feel like it is undermining credit unions’ effort to improve their members financial health, but the reality is that it may be supporting a reallocation of resources from lower impact, higher cost work to higher impact interventions that drive real change in members’ financial lives: helping members act, follow through, avoid irreversible mistakes, and connect one financial decision to the financial future they’re trying to achieve.

What your credit union can start doing today

1. Make your financial well-being offerings legible to AI. State clearly who each program serves, at what life stages, what a session covers, and how to reach it. A model can only surface what it can parse, and language about financial well-being programs tends to be vague.
2. Identify how you can include members whose needs do not fit neatly into what you currently offer. Most credit union guidance is built for one segment: members early in their financial journey, and that’s mission-critical work that should continue. However, you also want to support the member who has a 401k rollover question, is planning for a rental property, or has a stock-vesting decision—the content on your website needs to clearly reflect that you can also support those needs.
3. Identify the friction members experience when trying to get advice from you. The 24/7 accessibility of AI is setting a new precedent for ease of accessing financial content. There is no expectation that you meet this, but it’s a good nudge to access how much friction is in your current process. Pull information such as: how many members start the request, how many complete a session, how long between the two, and what the median wait looks like. Consider where there’s opportunity to make support more accessible, especially on more complex subjects where members still want a person to speak with.
4. Train staff to be the second opinion. Your frontline will soon spend its days responding to "ChatGPT told me...". Some of it will be good advice, some of it will not. Start by collecting what members actually walk in saying, and then use those examples to build staff awareness and response guidance. On this recommendation, it’s worth reinforcing that your staff should also have access to trusted AI tools to be able to quickly find and validate information.
5. Focus on financial outcomes not just on providing the answers. ChatGPT will give a better generic answer about any financial concept, and it will do it in 15 different ways if the members asks it to. You’re not going to out-answer the model, but your credit union can be the only one who says given your actual financial picture, here’s what this would do, and then help the member execute it.

The bottom line

Members will keep asking AI about their money, and most of the time, they will get reasonable answers. Treating AI as a competitor for financial information obscures the different roles you stand to play in a member’s life and may lead your credit union to make less informed financial well-being investments. This may be an opportunity for your credit union to focus less on providing basic information and more on your role as trusted partners, spending more time helping members move forward, achieve their financial goals, and improve their well-being.

—YP