At McKinsey’s 22nd Global Digital Banking Conference in Barcelona last week, 300 executives representing 220 financial services institutions from around the world gathered to discuss the issues at the center of digital banking, including agentic AI and a new era of geopolitical volatility. The event’s very name—the Global Digital Banking Conference—highlights its broadening relevance: The overriding message from leaders was that today, every bank should be a digital bank, because those that aren’t are at risk of getting left behind.
Agentic AI was on everyone’s mind. If anything, we may be underestimating AI’s impact in the short term, said McKinsey Senior Partner Harald Kube, adding that multiple banks are now using AI agents at scale. One of them, 82-year-old Latin American bank Bradesco, which presented during the conference, is prioritizing agentic use cases that assist in fraud prevention and that act as personal customer concierges. Bradesco’s CIO, Edilson Dias dos Reis, said the bank’s AI pursuits have freed up employee capacity by 17 percent and reduced lead times by 22 percent. In bringing agentic AI into the fold, every financial services player—start-ups and legacy institutions alike—can transform its value pools, including revenue and cost structures, customer experience, and job families and operating models.
Geopolitics, too, was at the forefront of discussion. Keynote speaker Arancha González, dean of the Paris School of International Affairs (PSIA) at Sciences Po, pointed out that geopolitical disruptions are reshaping trade, technology, and finance. According to González, three factors—security, emerging resource and industrial battlegrounds, and “transactionalism”—are testing globalization’s staying power. Even so, she doesn’t see an end to globalization: “Undoing interdependence has a huge cost that people have to bear.”
In addition to macro-level challenges, financial services leaders must confront difficulties at the organizational level. Even though banking is the largest industry in the world and spends the highest proportion of revenues across sectors on tech, poor labor productivity beleaguers financial institutions.
Following are themes we heard at this year’s Global Digital Banking Conference.
Agents are an increasingly important focus for banks. Gen-AI-enabled “agents” or “agentic systems” refer to digital systems that can independently interact with other agents as well as people and can be thought of like skilled virtual coworkers. These systems can act as “the completion of a puzzle,” McKinsey Senior Partner Leorizio D’Aversa pointed out. “It all started with automation and AI models, then gen AI to complement those insights,” he added. “Now, it is all packaged into agentic AI delivery, but agentic alone would not produce the same benefits without all the other ingredients.”
For banks to pursue agentic AI, however, they must establish the enablers that also allow for gen AI adoption to happen at scale—for example, a deep bench of tech talent, robust risk guardrails, and an updated operating model. To that end, McKinsey Senior Partner Gökhan Sari predicted that next year’s theme will be tech talent, particularly the organizational impact of agentic AI on people.
Gen AI remains a top priority. Of course, you can’t talk about agentic AI without also discussing gen AI. Continuing from last year’s conference, gen AI remains a top priority, and sector leaders are scaling their AI experiments fast. Just under half of respondents in a recent McKinsey survey said they are using gen AI regularly (46 percent today versus 34 percent in 2023), and 37 percent of respondents said they expect to increase their gen AI investments by more than 20 percent in the next year. The takeaway: Anyone still in the experimentation phase may very well be far behind their peers.
According to Alexandra Mousavizadeh, the CEO of banking research firm Evident and one of this year’s conference speakers, “breakaway” banks in AI adoption have emerged in the last year, increasing their AI adoption at double the rate of the average bank. (To be sure, even the average bank has been quick on the AI uptake.) That means banks can afford to be fast followers, but they can’t afford to be slow followers, Mousavizadeh said.
While banks are using agentic AI for fraud prevention, as personal customer concierges, or to enhance pricing algorithms, as in Bradesco’s case, gen AI use cases with the greatest potential value for banks include those related to corporate and retail banking as well as software engineering. Double-clicking into retail banking, virtual assistants, or AI copilots, are making contact centers more productive. Several banking players are pursuing these gen AI use cases today but could go further in applying gen AI to their customer operations and research and development. To successfully adopt these use cases, banks must also adopt end-to-end and systematic transformations.
The appetite for cloud keeps growing. More than half of banks now have mature cloud programs, and respondents in a recent McKinsey European Digital Banking survey said they plan to double the share of applications that are on the cloud in the next three years (from between 30 to 40 percent today to up to 70 percent in three years). In part, this may be because some of the bottlenecks that banks encountered in cloud adoption have been unblocked. Even so, banks face talent gaps that threaten their ability to go deeper: 88 percent of executives say they perceive the lack of cloud talent as the biggest obstacle to delivering a successful cloud program. Additionally, banks may find themselves stuck with a single cloud provider, which makes it harder to be agile.
Speaking of agility, there is perhaps no better case study of nimbleness in cloud migration than that of the Ukrainian bank Privatbank, which serves 24 percent of the Ukrainian banking market and was one of the conference’s presenters. PrivatBank’s technical infrastructure was highly concentrated in one data center. Within the first three days of the outbreak of the war in Ukraine, missile attacks occurred close to these data centers, prompting an emergency cloud migration. The migration, which would normally take about three years to complete, was executed in 45 days. “Typically, we talk about how we should first fix things in cloud, make them better, and then transition, so this transition was quite impressive,” said McKinsey Senior Partner Gökhan Sari. Would that approach be replicable for banks not facing a similar crisis? “Yes, definitely,” he said.
Targeting three segments can help banks achieve gains. Banks have meaningful opportunities to take market share in the areas of retail banking, with a focus on Gen Z consumers and affluent consumers; small and medium-size enterprises (SMEs); and corporate banking.
Take SMEs, which are burdened by administrative tasks and often use several different financial service providers (more than 50 percent have more than one bank). In Latin America, SMEs, which McKinsey Partner Paola Castilho called the engine of Latin American economies, represent the fastest-growing segment in financial services over the last several years. The unlock for banks to work with SMEs is scalability and digitization, she said.
Digital trust is now a business imperative. Digital trust—a bank’s ability to manage digital and technological risk—is a key differentiator that is also paying off in the markets. Banks that meet these definition of trust returned 7.8 times greater CAGR from 2017 to 2024 than nontrusted banks. Furthermore, trust is related to customer loyalty: Only 18 percent of customers said they were willing to shift away from a bank they trusted.
Talking about tech is correlated to winning in tech. Here’s where walking the talk translates to the bottom line. A McKinsey analysis found that banks that speak publicly about their tech pursuits are often better at connecting them to outcomes investors care about—D’Aversa said that connecting tech investments to outcomes investors care about is now “a differentiating element for performance.” The disclosure of tech KPIs, such as percent of digital sales and tech debt, has increased 150 percent from 2021 to 2024. What that means is that the banks communicating their progress in digital are also more likely to be better at measuring the ROI of their AI use cases, for example.
Given all the uncertainty in the year ahead, digital banking players should not ask themselves “whether they can navigate this period,” said McKinsey Europe Managing Partner Tunde Olanrewaju. “Rather, it’s whether you’ll make the right choices to do so successfully.”
Copyright © 2025 McKinsey & Company. All rights reserved.