Celent’s Technology Trends Previsory, Retail Banking 2027 Edition is now available online and Celent clients can access the full report via this link.

The core of the report covers five technology trends for retail banking in 2027:

Access deeper industry intelligence

Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.

Find out more

  • The new era of customer engagement
  • How the customer proposition connects to the core banking platform
  • Decision intelligence in lending and deposits
  • The shift to payments orchestration, and
  • AI as a catalyst for legacy modernisation.

Technology trends in banking they don’t happen in a vacuum. They’re shaped by forces outside the bank and inside it, globally and locally.

Opening Celent’s retail banking previsory, Zil Bareisis, Banking and Payments Director at Celent, said: “The economy has gotten tougher, and we’ve got to acknowledge that. The IMF expects global growth of just 3.4% in 2027. Inflation is back above 3% in the US the UK, and the euro area, partly because of energy prices. Households are under strain. US household debt is now $18.8tn.

“For banks, this means a renewed fight for deposits. And of course, while high interest rates for longer create higher credit risk, they also represent an opportunity for banks to increase net interest margin.”

What is a Previsory?

Part preview, part advisory, Celent’s annual Previsory series delivers a forward look at the technologies redefining financial services and how to respond to them.

Last year, Celent organised the Previsories in Banking and Payments around three key technology trends and imperatives, which also provided the backbone of its research agenda through 2026.

Overarching technology drivers

Bareisis noted the two overarching technology drivers impacting banking today, with AI moving faster than anything previously tracked. It’s not only changing how banks and vendors build software, how they run operations and how they run the business. It also changes how consumers are going to shop, pay and interact with their finances. And while costs and risks are still uncertain, the gap between banks that scale AI and banks that don’t will widen.

At the same time, new forms of money are moving from the experimental to infrastructure. Stablecoins and tokenised deposits now compete for the same balances. Regulations, like the Genius Act in the US and MiCA in Europe are creating clarity, and crypto-native firms are getting bank charters.

Bareisis added: “Retail banks need a plan to defend the deposit relationship. Finally, institution resilience is increasingly under attack by AI. This applies both to the scale and precision of deception attacks, but also increasingly exploiting value chains between financial ecosystem participants, not just at the perimeter of a bank, but also within those companies. And nobody wants to be the weakest link.”

A new era of customer engagement

Conversational, AI-assisted, human backstopped service is becoming the norm. The first one is a new era of customer engagement. Secondly, the customer proposition meets the core. Personalisation depends on the experience layer at the top of the stack, but also on the core system beneath it.

Added Bareisis: “You can’t really fix one without the other. Third: Decision intelligence is now embedded in lending and deposits. AI-based credit, fraud, and pricing decisions converge across the whole customer lifecycle. Four: A shift to payments orchestration. Consumer payments are becoming invisible and fragmented, so banks need to orchestrate them, not just process them.

“And then five, AI as the modernisation catalyst. AI is cutting the cost, time, and risk of escaping legacy cores to improve resilience.”

AI is becoming a part of life: Banks must meet the new engagement expectations

AI is really changing how people expect to interact with financial institutions and how they structure critical workflows on the job.

Michael Bernard, Celent principal analyst said that banks will need to address and support the new ways that customers and employees expect to find information, to collaborate, and to get things done.

“What’s happening is that consumers, increasingly are experimenting with AI assistance for personal financial advice, among a host of other lifestyle-related queries. At the same time, leading banks have begun introducing very capable digital and voice AI assistants within their digital channels and call centres.

“Bank employees are increasingly working with AI coworkers or tools, which prompt entire departments to rethink their longstanding operations models. Finally, and really importantly, we see major regulatory authorities becoming more supportive of automated relationship banking activity. That is really changing the game for relationship bankers and allowing banks to responsibly scale high-quality financial services to the mass markets.

“But how much appetite is there amongst consumers for AI services from their banks? We’ve done additional research and revealed that while consumers have become fairly enthusiastic about AI-powered advice and guidance from their bank, they’re far less likely, at least today, to engage in automated money movement, if it was offered by the bank.”

Show me, teach me – do not do it for me – just yet

Bernard said this was a case of show me and teach me, but not necessarily do it for me – just yet. For banks, the winning formula will be to introduce the AI-assisted tools that customers do want. That will require a clean execution in order to build trust. Moreover, banks must retain human support models as a backstop to retain that trust and loyalty.

Bernard flagged up recent Celent consumer research focussed on personalisation. Specifically, banks’ ability or otherwise to tailor products and services to reflect a customer’s total relationship with the institution has become one of the critical features that consumer and small business customers are seeking from their bank.

Added Bernard: “Banks will be constrained to deliver these types of experiences with their existing legacy core and product platforms.”

Celent’s take is that the winning bank model today is AI-assisted and human-backstopped, starting with guidance, discovery and support.

How the customer proposition connects to the core banking platform

Daniel Mayo, principal analyst, Celent, highlighted the firm’s large-scale Dimensions survey which revealed that around two-thirds of retail banks globally state that providing personalised customer experiences is a core priority for their institution.

“But where does that personalisation actually happen? Engagement, apps, assistants, advisers, third party AI agents – knows the moment. Intelligence and decisioning, that is customer data, models, next-best-actions and pricing – knows the customer. The product and pricing engine, a configurable product factory, rules, bundles and rewards, defines what can be offered. And finally, the core ledger and servicing – so real-time accounts, limits, interest rates and fees – all determine what can be executed.”

‘Where legacy gets in the way’: Daniel Mayo

Mayo gave a number of examples of legacy issues, including:

  • Product parameters hard coded and changes through IT release cycles;
  • Batch processing blocks real-time pricing, limits and rewards;
  • Customer data split by product silo, with no single view to personalise against;
  • Every new variant adds cost, complexity and operational risk, and
  • Vendor roadmaps set the pace of change

Celent’s take

Banks can personalise the message at the glass for a while but real proposition differentiation needs a configurable product layer and a real-time core.

To close the gap with customer expectations requires banks to fix the core platform deficiencies, not just focus on the front end. Mayo summarised five key steps:

  • Separate product from ledger;
  • Make data real time and event driven;
  • Personalise within guardrails
  • Modernise progressively, with intent
  • Manage proposition agility as a KPI

Mayo said that bottom line can be summarised thus: The ability to personalise is an architecture decision as much as a marketing one.

Decision intelligence embeds in lending and deposits

Celent principal analyst Craig Focardi noted that one of the most important shifts happening in banking right now is how financial institutions make decisions about their customers. Focardi stressed that the way many financial institutions do it today is holding them back.

“Banks need to replace siloed rules-based decisions with decision intelligence that runs within and across the entire customer engagement lifecycle for lending and for deposits. This is required to integrate AI and to provide a better customer experience and better offers.”

Looking first off at data, Focardi said that most credit and deposit decisions still lean heavily on credit bureau data. So, if a new customer opens a current or savings account or applies for a credit card account, the financial institution will obtain a FICO score.

And that works just fine for customers with long, clean credit histories and it is okay for simple account opening processes. But what of thin file customers, gig economy workers, or more sophisticated consumers, with different types of income? Customer income that arrives irregularly is not captured as well by a credit bureau view. Focardi said that the mass market and mass affluent customers can also benefit from greater personalisation and more information about their financial profile.

He said that the foundational answer is broader data, cash flow analysis, open banking data, and buy now, pay later activity, all used with the customer’s consent. The result is a richer picture of who someone is and what they need. Second, Focardi noted the need to consider overlapping systems. In most financial institutions, deposits, cards, mortgages, and auto loans each run their own origination IT stack with their own rules.

That means multiple overlapping technologies – and so the bank may see the same customer in multiple ways. Decision intelligence moves to one decisioning platform. That is one that is shared across products, channels and the client lifecycle, from onboarding to account management and collections.

Added Focardi: “Joint optimisation entails a customer-level relationship-based product and pricing decided together, along with credit risk and fraud prevention. And it means building trust through personalisation. Finally, governed probabilistic AI: that means machine learning and generative tools, with explainability, monitoring and human oversight.

Focardi summarised Celent’s take: Decisions should follow the customer, not the product. And so, pre-approved limits, dynamic pricing and proactive retention offers on deposits all result from one view of risk and value. The end result is that the customer trusts that they are getting more value.

Key retail payment themes

Bareisis turned to payments – the backbone of banking services and what customers expect from their banks. However, it is one notable area of customer dissatisfaction. When Celent asked consumers for examples of ways in which they expect more from their banks, payments consistently came back as one of the top priorities.

The six big themes for 2027

  • Cards and A2A payments increasingly co-exist
  • Global expectations, fragmented reality;
  • New money and payments infrastructures;
  • Relentless battle against fraud and scams
  • Wallets and platforms control the customer experience
  • Is agentic commerce getting ready to take off?

The shift to payments orchestration

Bareisis noted that banks traditionally provide current accounts, sell cards, process payments, fight fraud, show transactions and optimise individual payment products.

The emerging view-payments orchestration-means that the banks of tomorrow must orchestrate payment methods. In addition, they must connect different forms of money and must authorise AI agents as well as humans. They must not just fight fraud but predict fraud and financial risk. And the bank of tomorrow most use payments data intelligently and treat payments as a business.

Bareisis summarised Celent’s take as follows: banks must move from processing to orchestration. That means addressing the legacy challenge and build an adaptable, multi-payments platform.

AI as the catalyst for legacy modernisation

A key takeaway of the webinar is the Celent finding that legacy modernisation is growing in importance as a leading retail banking IT priority. According to the Celent Dimensions Surveys, this ranked only 6th in importance in 2024 out of eight IT spending drivers. In 2025 it rose to 4th in order of importance as a priority. In 2026, it ranks as the second most important priority.

Daniel Mayo said that the need for modernisation itself has been a long-standing issue in banking. There are, said Mayo, two forces catalysing this shift in importance. The first force is a push one, which is operational resilience. This has been particularly driven by banking regulators, which expect banks to deal with outages, obsolescence, and the loss of staff who knows these systems. Significantly, in addition to preparing for issues when they occur, there is a shift towards forcing institutions to proactively tackle potential resilience risks. The second force is driven by the desire of institutions to be able to incorporate AI effectively.

Mayo highlighted legacy core issues that limit real-time data, access to business logic and deficient process granularity of the level required for agentic AI to really deliver a step-changing performance. Banks are taking two paths here, said Mayo.

Around 20% plan a full system replacement. This is, more often than not, banks in Asia Pacific in the MEA region.

Said Mayo: “On the other side, just over 30% plan major modernisation of their existing systems, with this path more common in North America and Europe. With banks here looking to progressively modernise their platforms, but alongside its role as a driver for modernisation, AI will also become a critical enabler for modernisation across both paths, improving the business case by reducing cost, timeframes, and transformation risks. This is actually across the whole lifecycle: So, knowledge extraction, design, requirements, testing, and build.

“While use of AI for development and build has become pretty well established, one of the notable pain points in most modernisations is truly understanding existing systems in depth to be able to map and identify gaps as banks either rearchitect or move on to new platforms. Mayo concluded: “Celent expects 2027 to really see AI-driven modernisation gain mainstream adoption.”

The common thread: Trust becomes the moat

Bareisis stressed that the five key trends for 2027 all point in the same way. Namely, values moving from the interface the bank owns to the intelligence and orchestration behind it.

“And what ties them together? In our view, it’s a single word: trust. We think that because of all the forces that are taking place, customer inertia is going to erode. For decades, loyalty has been the quiet advantage of incumbent banks. Customers stayed because moving was a hassle. That’s changing. With AI assistance, they now can reduce attention friction. Wallets, real-time rails and stablecoins make it easy to move payments and balances.

“So, when inertia goes, trust is what’s left. Once AI acts on the customer’s behalf, the questions get sharper. Who is accountable? Who can be trusted with that identity, consent, and money? That means AI-assisted engagement with a human backstop, decisions that are governed and explainable, and clear rules on liability and payments. It even means agent access policies with a kill switch, if necessary.

The winning institutions will earn trust one real-time decision at a time. And modern platforms actually make it possible. None of this works on batch systems, so personalisation, instant decisions, and AI agents all need real-time data and configurable products, which legacy core struggle to deliver. The good news is that AI lowers the cost of fixing that.

Said Bareisis: “We think that 40% to 80% savings in cost and time on modernisation are quite realistic. So our concluding thesis is this: retail banks win by becoming the trusted orchestrators of customer context, decisions, money, and AI agents.”