The next major competitor to the banking app may not be another bank, fintech or digital wallet. It could be the AI assistant people already use to organise the rest of their lives. Reports that Anthropic is exploring ways for Claude to connect with personal financial information point towards something considerably bigger than another budgeting tool. If AI assistants gain permissioned access to bank accounts, transactions and other financial data, they could become the place where customers understand their finances, make decisions and eventually instruct financial services to act. The bank may still hold the money and process the transactions, but the intelligence sitting between the customer and their financial life could increasingly belong to somebody else.
The Banking App Could Lose Its Most Valuable Role
Digital banking has spent more than a decade trying to become the customer’s financial home. Banks have redesigned mobile applications, added spending insights, budgeting tools, investment dashboards, subscription management and personalised recommendations because controlling the interface meant controlling a significant part of the customer relationship.
AI potentially changes that equation because a sufficiently capable assistant does not need to recreate the traditional banking application. It does not need a screen containing accounts, cards, transfers and dozens of menu options. It needs permission to access the underlying information and enough intelligence to understand what that information means.
Instead of opening a banking app, searching through transactions and manually working out whether a large purchase is affordable, a customer could simply ask whether spending $5,000 on a holiday next month would affect their savings target. Answering that properly requires considerably more than displaying an account balance. The AI would need to understand income, recurring expenses, credit-card payments, savings patterns and upcoming obligations, potentially across several financial institutions.
That is where this becomes strategically important. The institution storing the money does not necessarily have to be the institution explaining the money.
AI Could Finally Unlock What Open Banking Started
Open banking made financial data portable, but AI could make that portability considerably more useful. The first generation of open banking largely focused on connectivity, allowing authorised third parties to retrieve account information, aggregate accounts or initiate payments. The infrastructure was important, but consumers still generally had to interact with applications designed around predefined functions.
Generative AI removes much of that structure because customers no longer need to know which feature or menu they require. They can simply describe the financial problem they are trying to solve. A customer could ask why spending has increased this month, which recurring payments could be reduced, how much they could realistically save by December, whether they should repay a credit card or retain more cash, or what would happen to their finances if their rent increased by 10%.
The interface becomes a conversation while the complexity disappears underneath it. This could dramatically increase the value of financial-data connectivity because AI can potentially interpret thousands of transactions and relationships without requiring customers to navigate dashboards or understand how the underlying financial products work.
The Bigger Shift Is Financial Aggregation
There is another reason banks should pay attention. An individual bank usually sees only part of a customer’s financial life. Their salary may arrive at one bank, their mortgage could sit with another, investments might be held through a brokerage platform, their credit card could come from another institution and their pension or insurance products could exist somewhere else entirely.
An AI assistant connected across those services could potentially understand the customer more completely than any individual financial institution. That reverses one of banking’s traditional advantages. Banks have historically possessed extraordinary amounts of proprietary customer information, but in an open-finance environment the strategic advantage may gradually shift from who stores the data to who can assemble, interpret and act upon it.
An AI assistant could effectively become the customer’s financial intelligence layer. Once that happens, it can also begin questioning the institutions sitting underneath it. If a customer asks whether they are getting a good return on their savings, the useful answer is not simply the interest rate currently being paid. An intelligent assistant could potentially understand how much liquidity the customer actually needs, identify money that has remained idle and compare alternatives available elsewhere.
The bank holding the deposit would suddenly have software continuously evaluating whether that deposit should remain there.
Financial Advice Could Become Continuous
Traditional financial advice is largely episodic. Customers meet an adviser, conduct a financial-planning exercise or actively seek information when something changes. An AI connected to live financial information could operate very differently because it could continuously recognise that spending has increased, savings are falling behind target, an insurance premium has risen, a subscription is rarely used or excess cash has remained idle for months.
Financial guidance therefore begins moving from something customers actively request towards something that emerges continuously from their financial data. That could be enormously useful for consumers, but it creates an important strategic question for banks: who owns the moment when a financial need is discovered?
Banks already spend significant resources trying to identify those moments through analytics, customer segmentation and next-best-action systems. In an AI-mediated financial environment, however, an external assistant may recognise them first. A mortgage opportunity, investment requirement, refinancing opportunity or savings need might exist inside banking data, while the first platform capable of recognising it could increasingly be the AI layer rather than the financial institution holding the account.
This connects directly with a broader issue Finnoex explored in The Next Generation May Never Choose a Bank. If financial services increasingly become embedded, aggregated and mediated by intelligent software, consumers may care considerably less about which institution provides the underlying financial infrastructure.
The Real Transformation Begins When AI Can Act
Allowing an AI to read financial information is important, but allowing it to act on that information would fundamentally change the relationship.
The progression is relatively easy to imagine. First, the assistant can see transactions and balances. Then it can explain them and identify patterns. After that, it can recommend actions. Eventually, with the appropriate permissions and regulated infrastructure, it could potentially execute those instructions.
A customer might tell an AI to maintain a certain emergency balance, ensure credit cards are paid before interest is charged, move surplus cash into an appropriate savings product and request approval before making any transfer above a predetermined amount. At that point, the customer is no longer manually managing individual financial transactions. They are establishing objectives, permissions and boundaries while software manages financial activity within those boundaries.
This is where agentic AI could have particularly significant implications for banking. As Finnoex has explored in its coverage of AI agents and financial services, the industry is moving towards a world where software may increasingly perform tasks on behalf of customers rather than simply providing information to them.
The difference between an AI that tells someone they are spending too much and an AI authorised to restructure how their money is allocated is enormous. It introduces questions around authentication, liability, consent, fraud prevention, transaction limits and what happens when an autonomous decision produces an undesirable outcome.
Permission Could Become More Important Than Personalisation
Financial information is also among the most sensitive data consumers possess. Transaction histories can reveal where someone travels, what they buy, which services they use, how much they earn and significant details about their lifestyle. Giving an AI persistent access to that information requires a different level of trust from asking it to summarise a document or write an email.
The future model is therefore likely to depend heavily on permissioned connectivity rather than simply handing an AI banking credentials. Existing open-banking infrastructure already provides a foundation where customers can authenticate directly with their financial institution and authorise specific information to be shared with another service.
AI, however, introduces a more complicated permission problem because access is no longer necessarily binary. Customers may want an assistant to analyse transactions but prevent it from initiating payments. They may allow it to monitor savings but restrict access to investments. They might permit small transfers automatically while requiring explicit approval above a particular amount.
The competitive advantage in financial AI may therefore depend as much on permission architecture as model intelligence. The most trusted financial assistant may not be the one that knows everything about the customer, but the one that makes it exceptionally clear what it knows, why it needs that information and precisely what it is allowed to do with it.
Banks Could Become Infrastructure Behind Someone Else’s Intelligence
There is an uncomfortable strategic possibility underneath all of this for the banking industry. Banks may remain indispensable while becoming considerably less visible.
The regulated institution could continue holding deposits, extending credit, performing KYC, monitoring financial crime, settling payments and maintaining the financial ledger while an AI platform increasingly owns the customer’s everyday financial relationship. The bank remains essential infrastructure, but another company controls the interface through which the customer understands and manages their money.
That possibility resembles what has already happened elsewhere in technology, where valuable infrastructure increasingly operates invisibly beneath another company’s customer experience. AI could accelerate the same development in financial services because consumers may eventually care less about which application contains their account and more about which assistant understands what they are trying to achieve.
If consumers begin asking Claude, ChatGPT or another AI assistant questions about their finances before opening their banking application, the competitive battleground moves upward in the technology stack. Banks would no longer compete only against other banks for customer engagement; they could increasingly compete for relevance inside an intelligence layer deciding which financial product, provider or action best satisfies the customer’s objective.
Owning the Account May No Longer Mean Owning the Customer
Anthropic’s reported move towards connecting AI with personal financial information should therefore not be viewed simply as another technology company experimenting with a personal-finance feature. It points towards a potentially much larger transition in how consumers interact with the financial system.
Internet banking moved financial services from the branch to the browser. Mobile banking moved them from the browser to the smartphone. Open banking began separating financial data from the institution that originally held it. AI could now separate financial intelligence from the banking interface itself.
The bank can continue holding the account, the payment network can continue moving the money and specialist fintechs can continue providing individual products. But if an AI understands the customer’s broader financial position, identifies what requires attention, recommends what should happen and eventually coordinates the action, it could become the layer the customer considers to be their primary financial relationship.
For banks, that may be the deeper implication of giving AI access to financial accounts. The biggest risk is not that an AI assistant can see a customer’s bank balance. It is that, eventually, the customer may stop needing to open their banking app to understand what to do with it.
What it means for the industry
- Banking apps face a different kind of competition: AI assistants could increasingly become the interface through which consumers understand and manage their financial lives.
- Open banking becomes more powerful when combined with AI: Connectivity provides access to financial information, while AI provides the intelligence required to interpret it.
- Customer insight could move outside the bank: An AI connected across multiple institutions may eventually build a more complete financial picture than any individual provider.
- Financial products could become machine-selected: Banks may increasingly have to compete for selection by AI systems comparing products and services on behalf of customers.
- Permission architecture becomes critical: Consumers will need granular control over what an AI can see, analyse, recommend and eventually execute.
- Agentic finance changes the risk equation: Moving from financial insights to autonomous financial actions introduces entirely new questions around authentication, liability and control.
- Banks could become invisible infrastructure: Financial institutions may continue providing essential regulated services while AI platforms increasingly control customer interaction.
- Owning the account may no longer mean owning the relationship: The strategic advantage could move towards whichever platform best understands the customer’s complete financial life.
