Marco Li Mandri from ING argues the smartest place to start with AI isn’t flashy “new bank” business models, it’s improving what already exists. There’s a lot of value in augmenting core processes first, so teams can move faster, reduce friction, and make customers’ lives easier without rebuilding everything from scratch.
Li Mandri points to mortgages as a strong example as these journeys often slow down because of manual checks, document chasing, and handoffs between systems and suggests that agentic AI could significantly shorten the “time to yes”. In practical terms, this could mean delivering an approval within a day, which is a meaningful win for customers making time-sensitive, high-stakes decisions.
ING also highlights voice as another high-impact area and the value isn’t voice on its own, but what it enables in service: customers explaining an issue naturally and getting it resolved quickly, without menus, delays, or switching channels. When voice is connected to systems that can actually complete tasks (not just record them), it becomes a genuinely new kind of service experience.
A key theme in Li Mandri’s view is sequencing as before moving into new business models, banks should extract value from foundational improvements and “crack” everyday problems like speed, accuracy, and customer effort. That’s especially relevant in a regulated environment where trust, controls, and consistency matter as much as innovation.
Once those early use cases are proven, ING says the same approach can scale into more domains, including anti-money laundering (AML) and internal customer service. The message is pragmatic: start with real pain points and measurable outcomes, then expand, so transformation becomes a confident rollout, not a risky leap.