Banks have spent heavily on artificial intelligence, cloud infrastructure, real-time data, digital lending, automated risk management and personalised customer platforms, often promising that these investments will allow them to understand customers earlier, make decisions faster and respond more intelligently to changing conditions. The next interest rate cycle could provide one of the clearest tests of whether those promises translate into genuine operational advantage. The European Central Bank raised rates by 25 basis points in September as renewed inflation pressures complicated the economic outlook, while interest-rate uncertainty remains significant across major markets. For banks, another period of higher or more volatile rates would affect far more than lending margins. Depositor behaviour can change, borrower affordability can deteriorate, credit risks can emerge in unexpected places and funding decisions can become more complicated. Unlike previous cycles, however, banks now have an unprecedented amount of technology monitoring those changes. The question is whether that technology can help them respond before traditional indicators reveal that something has gone wrong.
Higher Rates Test More Than the Balance Sheet
Interest rates affect almost every important part of a bank. They influence the price of deposits and loans, customer demand for credit, borrower affordability, the value of financial assets and the economics of funding. The Basel Committee defines interest rate risk in the banking book as the potential impact of adverse rate movements on both bank capital and earnings, including changes to net interest income and the economic value of assets and liabilities.
The underlying risk is hardly new. What has changed is the technology available to understand it.
Banks today have considerably greater access to real-time customer behaviour, transaction data, automated credit monitoring and sophisticated modelling than they did during earlier rate cycles. Many institutions have also moved important analytical workloads into cloud environments, introduced machine learning into credit and fraud decisions and invested in platforms capable of connecting information that previously sat in separate systems.
Those investments should theoretically give banks greater visibility into how monetary conditions are affecting customers and the balance sheet. A higher-rate environment could determine whether they actually do.
Credit Models Are About to Meet Changing Borrower Behaviour
One of the clearest tests will come in lending. Traditional credit models are very good at analysing historical information, but an interest-rate shock is inherently forward-looking. A borrower who has never missed a payment may become substantially more vulnerable when refinancing occurs at a higher rate, while a previously healthy small business may experience margin pressure as financing, energy and supplier costs rise simultaneously.
This is where newer data and AI capabilities should begin to demonstrate their value. Banks increasingly have access to transaction patterns, cash-flow information and behavioural signals that can reveal deterioration before it appears in conventional arrears data. Instead of waiting for a customer to miss a payment, an institution could potentially identify declining account balances, growing reliance on revolving credit, changes in salary flows or deteriorating business cash flow much earlier.
AI is already being deployed by banks for creditworthiness assessment and risk management. The Bank for International Settlements has noted that banks’ credit-risk processes need to incorporate forward-looking information and has argued that supervisors increasingly need to consider how AI-driven structural changes affect the resilience of bank portfolios and business models.
The real test for AI credit models, therefore, may not be whether they approve loans more quickly during normal conditions. It may be whether they recognise changing risk faster when economic conditions stop looking like the data on which those models learned.
Personalisation Becomes More Important When Customers Are Under Pressure
Banks have spent years investing in personalisation and next-best-action technology. Much of the customer experience created by those investments has concentrated on recommending products, improving engagement or identifying sales opportunities.
A more difficult economic environment changes the purpose of personalisation.
If borrowing costs rise, a bank with a sufficiently developed customer-data infrastructure should potentially be able to distinguish between customers who remain financially comfortable and those beginning to experience pressure. The relevant next-best action for the second group may not be another product. It could be an early refinancing conversation, a change in repayment structure or intervention before financial difficulty becomes delinquency.
This is where the difference between having customer data and being able to act on it becomes important. Information may exist across current accounts, credit cards, mortgages, business accounts and payment activity, but if those systems cannot produce a coherent and timely view of the customer, the bank can still discover financial distress too late.
The interest-rate cycle could therefore become an unusually practical test of years of investment in customer-data platforms. Banks have repeatedly said that unified data enables them to understand customers better. Higher borrowing costs may reveal whether they can use that understanding when customers actually need it.
Deposit Technology Could Suddenly Become Strategic
Higher rates also change the other side of the balance sheet. Depositors have more incentive to compare returns, move excess cash and search for better-yielding alternatives when the difference between products becomes financially meaningful.
Digital banking makes that behaviour easier. Customers no longer need to visit another branch to move their deposits. In many markets, opening an alternative account and transferring funds can be completed from a smartphone.
This makes deposit analytics, pricing technology and customer intelligence more strategically important. Banks need to understand which customers are sensitive to rates, which deposits are likely to remain stable, where money is moving and how much they should pay to retain different categories of funding.
The lesson from the rapid banking stress of recent years is that digital finance can also accelerate behaviour during periods of uncertainty. The BIS has warned that digital claims can move or be redeemed faster than underlying assets can necessarily be sold or funded, potentially intensifying liquidity pressure during stress.
Technology therefore creates both sides of the equation. Digital banking can make deposits easier to acquire and manage, but it can also make them easier to move.
Loan Servicing May Become More Important Than Loan Origination
Much of fintech’s lending innovation has concentrated on the beginning of the credit journey. Digital onboarding, automated underwriting and instant decisioning have reduced the time required to apply for and receive financing.
A difficult rate cycle shifts attention further down the lifecycle.
When borrowers experience pressure, banks need technology capable of distinguishing temporary difficulty from deeper credit deterioration, prioritising cases and determining which intervention is appropriate. Collections systems, restructuring platforms, behavioural analytics and automated customer engagement can become just as strategically important as origination.
For SME banking, this may be particularly significant. Smaller businesses can experience changes in cash flow quickly and often have fewer financing alternatives than large corporates. The BIS warned in its 2026 Annual Economic Report that tighter credit conditions could expose vulnerabilities among smaller and middle-market firms, while renewed inflation or higher rates could contribute to a broader credit tightening.
A bank that can identify those pressures early has more options than one discovering them after repayment problems become visible. The technology advantage is therefore not simply faster credit decisions; it is potentially faster recognition that an existing credit decision needs attention.
Treasury Technology Will Face Its Own Test
Some of the most important technology during a rate cycle sits far away from the customer interface. Asset-liability management, liquidity forecasting, interest-rate risk measurement and treasury systems determine how effectively banks understand what changing rates are doing to their own balance sheets.
The Basel framework places considerable emphasis on behavioural and modelling assumptions, data integrity, model governance and reporting to management when assessing interest-rate risk. These requirements become particularly important when customer behaviour departs from historical assumptions.
A deposit traditionally considered stable may behave differently when customers can move money instantly and alternative savings products offer substantially higher returns. Mortgage prepayments can change. Loan demand can shift. Funding costs can respond differently across customer segments.
Better technology cannot remove those uncertainties, but it should allow banks to model more scenarios, update assumptions more frequently and give management better information about emerging exposures. In this environment, the quality of data feeding treasury and risk systems may matter as much as the sophistication of the models themselves.
AI Will Be Tested Outside Normal Conditions
The emerging use of AI adds another dimension that previous interest-rate cycles did not have at comparable scale.
Banks are deploying AI across lending, fraud, compliance, customer service and risk management. The IMF has observed that AI is increasingly moving trading, lending and financial analysis towards real-time decision-making, creating significant efficiency opportunities while also introducing questions around governance and correlated behaviour.
Systems that perform well during relatively predictable conditions may behave differently when economic relationships change. Borrower behaviour can shift, historical correlations can weaken and patterns that models treated as normal may stop being reliable.
That does not mean AI will necessarily perform badly during a rate shock. It means the cycle could provide a more meaningful assessment of AI than another benchmark measuring processing speed or prediction accuracy under familiar conditions.
Banks will need to know when models are becoming less reliable, when assumptions require recalibration and when human judgement should override automated recommendations. The BIS has recently made a similar point from the supervisory perspective: technology changes the speed and form in which familiar banking risks emerge, but it does not make traditional expertise in capital, liquidity, credit and interest-rate risk less important.
The most technologically advanced bank may therefore not be the one that automates the greatest number of decisions. It may be the one that recognises fastest when automation should be questioned.
The Value of Technology Will Be Measured in Response Time
This is ultimately what makes another interest-rate cycle so interesting from a banking-technology perspective. Banks have accumulated enormous amounts of data and invested in systems designed to turn that information into decisions. During stable economic conditions, the difference between receiving an important signal today and receiving it next week may not always be obvious.
During a period of rapid change, that difference can matter considerably.
A bank that identifies deposit migration earlier can adjust pricing or funding strategy sooner. A lender that detects deteriorating borrower cash flow earlier has more options for intervention. A treasury team receiving more accurate liquidity information can respond faster to changing assumptions, while risk teams with connected data can identify concentrations that might remain hidden inside individual business units.
Technology therefore does not eliminate the economic consequences of higher interest rates. It potentially changes the speed at which the institution understands those consequences.
That distinction matters because the financial industry has spent years describing AI, cloud, data platforms and automation primarily through efficiency. Faster processing, lower operating costs and better customer experiences have provided relatively straightforward business cases for technology investment.
A more difficult economic environment creates a tougher measure of value.
The next rate cycle could reveal whether banks have merely digitised their existing processes or actually built institutions capable of sensing changes and responding to them faster. If technology investment has worked as intended, banks should have better information, earlier warning signals and more options when conditions deteriorate.
If they do not, years of digital transformation may have made banking faster without necessarily making it more resilient.
What this means for the industry
- AI credit models will face a more meaningful test: Their value will increasingly depend on identifying changing borrower risk before it appears in traditional arrears and credit indicators.
- Customer data platforms need to demonstrate operational value: Higher rates could show whether unified customer data actually enables earlier intervention when financial circumstances deteriorate.
- Deposit technology becomes strategically important: Digital banking makes money easier to move, increasing the importance of deposit analytics, behavioural modelling and dynamic pricing.
- Servicing technology deserves greater attention: As borrower pressure increases, collections, restructuring and early-warning platforms may become as important as digital loan origination.
- Treasury and risk technology move closer to the centre: More volatile rates increase the importance of real-time liquidity information, asset-liability modelling and reliable behavioural assumptions.
- AI governance becomes more important during abnormal conditions: Banks need to understand when models trained on previous behaviour are becoming less reliable and when human intervention is required.
- Technology ultimately needs to improve response time: The competitive advantage may not come from having more data or more AI, but from recognising important changes earlier and acting before problems become significantly harder to manage.
