The Expert View: Leveraging AI and data to enhance credit risk strategies in commercial banking

As AI begins to move beyond proof-of-concept initiatives and into core banking systems, financial institutions across Europe are facing a more fundamental shift – one that goes beyond digital transformation and towards a genuinely AI-centric operating model.

 

This was the focus of a recent roundtable in Vienna, hosted by Business Reporter and sponsored by nCino. Bringing together senior executives from across the DACH region, the discussion quickly moved past the usual hype around generative AI to turn to a more practical question: what does it take to re-engineer core lending processes so that AI doesn’t just support decisions, but starts to play a more active role within them?

 

From experimentation to something more structured

 

For many in the room, the starting point was familiar. AI initiatives had typically begun two or three years ago as container pilots that were useful but limited in scope. What’s changed is the level of ambition.

 

“It’s not really the technical side that’s slowing things down anymore,” one participant said. “It’s getting the business to think differently – to design processes with these tools in mind from the outset.”

 

That shift – from isolated experimentation to something more co-ordinated and structural – came up repeatedly. In particular, there was a sense that real progress will depend on embedding AI across the full lending journey, rather than layering it onto existing workflows.

 

The reality of SME data

 

If there was one area where that ambition meets friction, it’s SME lending.

 

Several participants described it, bluntly, as a “data gap”. Unlike retail banking, where data is relatively standardised and plentiful, SME lending still relies on fragmented, inconsistent information.

 

That creates a challenge for AI – but also an opportunity.

 

“We’re starting to look at more contextual signals,” one executive explained. “Things like customer sentiment or external indicators. But the reality is, no underwriter can go through that manually at scale. That’s where AI starts to make sense.”

 

At the same time, there was recognition that bringing these new data sources into existing systems and making them usable remains far from straightforward.

 

Trust is still building

 

For all the progress being made, trust in AI-driven decision-making is still a sticking point.

 

Concerns about explainability, accountability and the reliability of large language models came up more than once. In particular, the idea of fully autonomous decision-making still feels like a step too far for many.

There wasn’t outright resistance but there was caution.

 

“AI can get you a long way,” one attendee noted, “but the final judgement still sits with the human.”

 

Anthony Morris, Chief Innovation Officer at nCino, made a similar point, but from a governance perspective: “This shift to an AI-centric operating model requires a level of governance and accountability that banks are already used to, but it now applies in a completely new way.”

 

In other words, the technology may be advancing quickly, but the frameworks around it still need to catch up.

 

A gradual move towards autonomy.

 

Looking ahead, there was broad agreement on the direction of travel even if views differed on the pace.

 

Most participants described a future where lending sits on more unified platforms, with AI supporting different types of workflows depending on complexity. At one end, fully automated processes for simpler lending. At the other, more complex cases where AI plays a supporting, analytical role.

 

Some see “agentic” systems – those capable of handling multi-step workflows – as the next logical step.

 

Others are less convinced, at least in the short term, pointing to regulatory constraints and the need to maintain confidence in decision-making.

 

Either way, the shift is already underway.

 

From digital transformation to AI transformation

 

If there was a single takeaway from the discussion, it was this: progress will come less from bold vision statements and more from solving specific, practical problems.

 

As one participant put it: “It’s not about doing AI for the sake of it. It’s about doing things better.”

 

That might mean earlier risk detection, better portfolio visibility or simply reducing the manual effort involved in existing processes.

 

The opportunity is clear. But so is the work required to get there, particularly when it comes to governance, data quality and rethinking systems that were never designed for this kind of model.

 

The shift from digital transformation to AI transformation is no longer theoretical. The question now is how quickly institutions can make it real.

 

---

To learn more, please visit: www.ncino.com and nCino EMEA Summit 2026