In partnership with

Sponsored deep dive essays are opportunities for me to work with specific companies on topics and issues important to them, me and bank readers. The sponsor provides input and information, but each piece is researched and written by me, and are my honest thoughts.

Thanks to MX for partnering with me on this piece. I’ve been thinking a lot about the tech moment banks are in, how we got here and what banks should do about it, and this was the perfect opportunity to write about it.

Agentic AI Opens Bankers’ Eyes to Open Banking

It’s 1 a.m., and one of your bank’s customers can’t sleep. They’re anxious about money, or something that costs money. Maybe it’s a tuition or hospital bill; maybe it's holiday or vacation spending; maybe it’s about buying a house or a car; maybe it’s about saving up for retirement.

They want to ask their question, but they’re worried about sounding stupid. They’re worried someone will judge them for having the question, or for being worried about money, or not having enough money to not be worried.

Instead, they open an app on their phone, type out their question and get an answer that helps them and reassures them that they’re not stupid for being worried. The response might ask some follow-up questions or offer to do some further research about potential options. It might even offer to act on their behalf.

Is the app your customer clicked on your bank’s? Did they talk to your bank’s AI agent, which uses customer-permissioned data to make tailored recommendations and offerings? Or did they go to an AI agent — Claude, Gemini, ChatGPT, Grok, Muse or Instinct — to get help?

AI agents that can assist consumers with personal finances are the latest incursion of AI tech firms on banks’ turf. Bank executives and stakeholders are concerned about how to respond. Some may see it as a threat to banks if consumers exercise agentic authority. That’s understandable, but it’s only half of the equation. Agentic AI offers financial institutions an opportunity to serve their customers better by understanding them and deepening the relationship.

How? By using open banking frameworks and capabilities.

Open banking is the structured and secure consumer-permissioned sharing of data. The specter of widespread personal finance agents gives financial institutions a powerful defensive rationale to implement open banking technology. It gives banks a way to offer greater data connectivity and add agentic AI capabilities. Best of all: They don’t have to wait for a rule to get started on adding the tech that will enable all of it.

The State of Open Banking

Open banking regulation in the U.S. has been in flux, and executives could be forgiven for being confused about where it stands. The Consumer Financial Protection Bureau published its open banking rule, Section 1033, in October 2024. It was immediately challenged in a lawsuit, which has taken many twists and turns since then. Most recently, the CFPB has indicated it will revise and republish the rule, which could happen soon.

The lawsuit and the prospect of a revised rule may have led executives to pause their implementation efforts. But the tech companies have shown that while rulemaking may be frozen, they’re certainly not waiting around — and neither are customers.

The rest of the world isn’t either. There are 61 countries with some form of open banking or open finance live or in development, according to a December 2025 article on Retail Banker International. The last two years have been especially important. The U.K.’s open banking ecosystem recorded more than 2 billion API calls and more than 15 million users in July 2025, signaling a turning point in maturity and adoption in the country, according to Open Banking Limited. And newer frameworks, like Canada's, will eventually extend to payment initiation, opening the door to agentic use cases like write access, which gives an AI system the authority to change, create, delete or commit data and operational states in external systems, rather than just reading or observing information.

In the US, the market has led the way with open banking, and a great deal of progress has already been made for consumers. The technology exists, the Financial Data Exchange recommends shared rules of the road and the largest banks are moving forward at pace, building out their open banking offerings for both retail and small business banking.

The lack of regulatory mandate may have persuaded some banks to pause their implementation of open banking tech or hold off altogether. But the market urgency still exists. AI agents are ready to assist your customers from their devices, all while disintermediating traditional financial institutions.

The Best Offense Is a Good Defense

Banks are staring down demographic and technological shifts. They have leaned on older customers who provide most of the assets and deposits and may have lower tech expectations. Their heirs likely won’t share those preferences.

Banks will need to offer products and services that resonate with younger customers to grow and remain relevant. About 60% of consumers in an MX survey say they expect their financial provider to leverage their data to personalize the experience, and a little over half of consumers expect greater levels of personalization from their finance-related mobile apps than they receive today.

While AI is making inroads with an increasingly comfortable consumer base, it’s not good enough to fully replace the role of a financial advisor — yet. Popular AI services provided accurate and complete answers to financial questions only 43% of the time, according to a comprehensive benchmarking study of 18 popular models from Saturn, an AI and technology firm, released in September.

The models made mistakes that included incorrect figures, outdated rules and omitting warnings or critical pieces of information. Free models performed worse than their paid counterparts; one free model even hallucinated a rule for student loans.

This could have real consequences for consumers using these models for financial advice. The study found that in the worst cases, someone who relied on the advice could lose “hundreds of thousands” of their money “or have other life-changing consequences.”

Of course, it’s possible that over time, the models will improve and the accuracy gap will close. But right now, banks actually have an advantage when it comes to leveraging open banking tech to create AI-powered experiences: actual, deterministic data on their customers. 

You see, when a customer asks Claude or ChatGPT a financial question, it’s usually about their money. “What should I do? What is the best loan for me? Help me save money.” These questions are best informed by a customer’s transaction data, not what the customer remembers or wants to believe about themselves and their money.

In an Agentic World, Friction Loses

Banks need to move away from a strategy that assumes their most profitable customers won’t want better technology offered by newer firms, and adopt a strategy that uses customer data and tech capabilities to better serve their customers.

So how can banks respond in an agentic world? MX, which partnered with me on this deep dive, provides open finance APIs, account aggregation and data-driven mobile banking solutions for financial institutions. It has developed a playbook for banks looking to implement open banking tech:

• Beef up security. The banking industry is seeing sophisticated fraud attempts and intrusions or attacks on infrastructure by groups ranging from lone actors to nation-states. In addition, agentic scraping could pose a real safety threat to banks and their customers. Security matters most when data is moving. Offering permissioned APIs is both a commercially sound decision and an ethical one.

• Get curious. AI isn’t the only tech that’s moved fast in the last couple of years. There are a number of companies that partner with banks to build and enable open banking tech. And banks can use newer tools to plug partners into their tech stack and configure solutions that resonate with their unique market areas.

• Build a data-in strategy. More than half of consumers in an MX survey said they would give their financial provider access to more of their data if they knew it would result in a better experience. Bringing in more of your customers’ data helps them, and you, get a full financial picture: outside checking accounts, credit cards, debt, savings and retirement plans.

• Craft how you’ll use that data. Use this customer-permissioned data and the full customer profile to offer the right next product — maybe that’s a higher-yield savings account or a car loan that fits within their financial picture. Open banking data can also help your institution identify customers who fit the core customer profile for special outreach and targeting.

• Install an embedded AI assistant. Banks can offer an AI agent that has appropriate security, guardrails and governance, built on deterministic data. Position your institution as a trusted partner for advice and recommendations, rather than forcing customers with questions to go elsewhere. Banks that build useful tools for their customers have a chance to keep them. Banks that don’t are hoping that their customers remain as fearful of agentic AI as they are.

• Embrace data portability. Customers want to share their data with their AI agents. Preventing them from moving their data out or charging data access fees creates friction that customers experience and could become an enormous disadvantage for institutions. Keep your customers by earning their business on the merits of your products and services

Don’t Be the Only One Waiting

AI insights, advice and agents turn open banking into a necessary defensive strategy for banks to maintain customers — but it's also a way to deepen relationships with existing customers and appeal to new ones. The longer banks hold off on implementing open banking tech and building personalized, proactive insights for customers, the greater the risk that they’ll be disintermediated. The cost of doing nothing rises each time AI tech companies put out a new model.

When your customer needs help at 1 a.m., a tech firm and its AI agent will be ready to assist them with their question and maybe even take an action for them. Make sure your bank is ready to help too.

Thanks so much for reading today’s special deep dive, in partnership with MX! As always, feel free to reach out with any thoughts or comments. - Kiah