.png)
Consumers are getting budgeting advice, debt reduction strategies, and financial planning guidance from general-purpose AI tools that have no relationship with the customer, no transaction history, and no fiduciary responsibility.
They are asking ChatGPT which credit card offers the best rewards, querying Perplexity about mortgage rates, and letting AI summarize complex financial products. And the bank - which holds all the data, all the history, and all the trust - is not in the conversation.
Last year, only 10% of consumers reported using AI to help manage their personal finances. Today, more than half say they use AI to aid their financial management decisions - with adoption highest among Gen Z (77%) and Millennials (72%), and growing steadily among Gen X (49%) and Boomers (30%), according to the 2025 TD Bank AI Insights Report.
The pattern holds globally. The EY Global AI Sentiment Survey of 18,000 people across 23 countries found that 49% of consumers had already used AI to support savings and investment decisions in the past six months alone - with adoption accelerating as consumers find practical use cases beyond experimentation.
In an episode of the Banking Reinvented podcast, Theodora Lau spoke about what this means for banks. She has spent a decade focused on the intersection of technology, money, and consumer well-being. She is the author of Banking on (Artificial) Intelligence, and host of One Vision Podcast.
"If consumers are not going to [banks] and they're going to the tools, what does that mean? Are they getting more value from these tools than they're getting from their banks and advisors?" she asked during her conversation with host Tim Rutten.
That is the question incumbent banks are not asking themselves urgently enough.
Banks missing the conversation is not a UX or tech problem
Banks tend to frame customer drift as a UX issue, but a better app and faster onboarding cannot alone reverse this.
Theodora's argument is that the issue is structural. Banks are optimizing for operational efficiency while consumers are demanding personalization, protection, and genuine financial guidance across the complexity of their lives.
"Financial services is not just about an individual. It's about a household. It's about multiple financial and non-financial relationships all coupled together."
She describes the sandwich generation - people in their 40s and 50s simultaneously managing their own retirement, their teenagers' college costs, aging parents, business cashflow, and healthcare expenses.
No checking account feature and no product refresh solves that. What consumers need is an institution that understands their full financial reality and can help them navigate it proactively, not reactively.
That is precisely what AI tools are beginning to offer, however imperfectly. As AI rapidly shifts from experimentation to everyday expectation, users are embracing its benefits while becoming more deliberate about where - and under what conditions - they want it applied.
The shift is not just about information but also about initiative. Consumers are moving from seeking static accounts to expecting intelligent assistance that acts on their behalf.
What does losing the first conversation with potential customers to AI mean for banks?
Losing the first conversation means losing the signal. It means not knowing what the customer needed, what question they were asking, and what decision they were weighing. And without that signal, the bank cannot be there at the moment that matters.
"If you're not getting the signals, you won't be able to be there at the moment they need you," Theodora adds.
And there is a second consequence that is harder to see. When a consumer asks ChatGPT about mortgage rates and gets a useful answer, they do not just get information; they also get a habit. They learn that this is where financial questions get answered. The next question goes there too, and the one after that.
The bank never gets a chance to demonstrate its data advantage because it was never in the conversation to begin with.
Continue reading: Conversational banking AI: OpenAI's move changes the game
Consumers are ready for AI-assisted banking, not for AI that replaces banks
There is a notable nuance here that banks should not ignore.
Despite accelerated adoption, trust is developing more gradually: only 18% of survey respondents say they would trust AI to make financial recommendations on its own.
According to the same survey, the vast majority of consumers - 90% - say they trust personal relationships, and 85% trust financial institutions. Nearly half (48%) say human review of AI-generated guidance would increase their confidence. This gives banks a trust advantage that they need to sustain.
Consumers are using AI tools but they are not fully trusting them. They want someone to verify the answer, and the human accountability behind the recommendation.
Consumers show the most comfort with AI being used behind the scenes - roughly two-thirds are comfortable with AI-powered product recommendations and fraud detection. Nearly half are open to AI banking assistants that help with proactive tasks, from paying bills to setting alerts.
That distinction matters more than most banks realize. The 48% who say human review would increase their confidence are not asking for a human in every loop. They are asking for accountability behind the recommendation. They want to know that someone is responsible for the answer - that if it is wrong, there is an institution that stands behind it.
General-purpose AI tools cannot offer that, but banks can. That is a structural advantage sitting unused.
Banks that understand this - and move to embed intelligent, proactive guidance inside the banking relationship - have an advantage no general-purpose AI tool can match: actual transaction data, actual financial history, and a trust relationship that has been built over years.
But that advantage expires. Every month that goes by where the bank is not showing up in these conversations is a month where the habit of going elsewhere becomes more entrenched.
From static accounts to intelligent assistance
There is a version of banking that consumers are beginning to sense is possible, even if they cannot fully articulate it yet. It looks like this:
- A bank that sees the car purchase forming in their spending patterns before they have asked about financing.
- A bank that connects the retirement account, the college savings, the business cashflow, and the healthcare costs into a picture that resembles their actual financial life rather than a list of separate products.
- A bank that notices the cash flow squeeze building and says something before it becomes a crisis.
That version of banking exists in pieces. What is missing is the architecture to bring it together.
Theodora uses a navigation analogy that captures this well. In life, we go from point A to point B, but this rarely happens in a straight line. The routes change constantly - new constraints appear, priorities shift, family circumstances evolve. What users need is something that adapts with you in real time, not something that shows them where they were last time you checked in.
The gap between what banks currently offer and what consumers are starting to expect is not primarily a product gap or a feature gap. It is an architecture gap preventing banks from acting proactively on a customer's behalf. A bank needs a unified, real-time view of that customer's financial state to do this, which spans every account, product, and life stage, and updates continuously as circumstances change.
Most banks do not have that today. They have data, but that data lives in disconnected systems that were built to serve specific products rather than to understand specific people.
That structural reality is why general-purpose AI tools have been able to move into this space at all. They have no access to transaction history, no fiduciary responsibility, and no real knowledge of the customer's situation. But they show up when a question is asked, they engage with it seriously, and they try to help.
What "serving the customer" means now
For decades, banks held a position that did not need defending. They were trusted, regulated, and established. No challenger seriously contested that.
Theodora argues that position no longer protects them.
"If you're not solving the questions and the problems that consumers have nowadays, why would they stay with you?" she challenges.
AI-native challengers and agile fintechs are responding to this by building from the customer outward, beginning with services that start with a need, not a product. They are hyper-personalized, fast, and indifferent to legacy reputation.
She argues that brand loyalty is not a strategy. It is a legacy asset that depreciates if it is not reinforced by genuine value.
She pointed to Nubank as an example of what getting this right looks like. The Brazilian neobank amassed over 100 million users by using technology to help consumers think about saving, making money stretch, and supporting the rest of the people in their family.
They built from the need inward, not the product shelf outward.
The contrast with incumbent banks is structural, which still mostly design for accounts not for lives. That is a mindset gap that is harder to close than a technology gap.
The habit is already forming elsewhere
49% of global consumers have already used AI to support savings and investment decisions in the past six months, according to the EY Global AI Sentiment Survey of 18,000 people across 23 countries - with adoption accelerating as consumers find practical use cases.
Consumers are not waiting for banks to figure this out. They are already building habits elsewhere, and the 55% figure will be higher next year.
That shift - from product-centric to genuinely customer-centric - is the real transformation, and the window to lead it is closing.
The banks that move now - embedding proactive, AI-driven guidance inside the relationship they already own - will be the ones customers do not need to leave.
