Exclusive: Celonis global banking head says AI rollout hinges on process intelligence

Banks around the globe are under mounting pressure to modernise, but the path forward is proving more complex than simply deploying new technologies.

Chris Johnston, SVP, Head of Global Banking at Celonis, said financial institutions are increasingly treating process intelligence as a foundational step in their artificial intelligence strategies, following months of experimentation with large language models and enterprise tools.

From cyber threats and regulatory scrutiny to the operational realities of integrating AI, financial institutions are being forced to reassess how they function at a foundational level. Increasingly, that reassessment is centred on understanding internal processes before layering in new capabilities.

Since joining Celonis nine months ago, Johnston has spoken with more than 50 banking executives across regions. Those conversations, he suggests, reveal a sector grappling with similar structural challenges, regardless of geography or business model.

"There's a lot of commonality across the global banks," said Johnston. Adding that all of them are looking for ways to improve things like onboarding, the customer service journey, trade and product operations, payments, and compliance.

"These desires are not new. However, the promise of reimagining these domains with AI at enterprise scale is."

That shift has exposed a structural issue across the sector. Banks have invested heavily in AI capabilities but lack a clear understanding of how their internal processes operate at scale. As a result, process intelligence, the mapping and analysis of how work flows across systems, is becoming what Johnston described as "phase zero" of enterprise AI deployment.

"In order to do that effectively, they're all coming to the realisation that the first thing they have to do is figure out how the business runs today. So they can figure out what they want to automate or put agents on, what they want to redesign and what they want to get rid of," said Johnston.

Onboarding, in particular, has become a competitive battleground. Delays or friction in these processes can directly affect revenue, especially where prospective customers abandon applications before completion. 

Celonis positions its platform as a way to create a digital representation of how processes actually run inside a bank, often referred to as a digital twin.

Historically, process mining has been used as a retrospective diagnostic tool. Johnston said that it is shifting toward continuous monitoring and orchestration.

"We absolutely talk about creating a digital twin, oftentimes in the bounds of the domain that we're focusing on - the ability for the banks to then use Celonis to mine the processes to use our orchestration capabilities, to create actions, to smooth out friction points, lay a composable light architecture over top of the existing systems."

The core issue, according to Johnston, is not access to AI technology but the ability to control and validate its outputs within highly regulated environments.

This uncertainty is driving caution in deploying enterprise AI across many organisations, with some executives saying they are unwilling to "unleash" many of the AI capabilities they have built or acquired until they have a better understanding of the implications and the highest-value initiatives.

"We're told that some financial institutions have chosen not to turn on the full breadth of their licensing agreements with AI companies, because they don't have confidence that they're going to get the benefit of that horsepower being unleashed in the organisation yet," Johnston said. "They don't have that process intelligence, digital twin capability, and they don't know exactly what needs to be optimised."

This common issue has led some banks to prioritise process mapping and monitoring capabilities before scaling AI initiatives, regardless of regional regulatory differences.

Johnston argued that the rise of enterprise AI is acting as a forcing function for operational discipline in banking, a sector that has historically been less constrained by efficiency pressures than industries such as manufacturing.

Banks are no longer looking for isolated improvements, but for platforms that can support a fundamental rethinking of their operations and provide the oversight needed to manage increasingly complex, hybrid environments of human and machine-driven work.

"What [banks] are really looking for right now is setting themselves up for the long term to reap the benefits and promise of the agentic future," Johnston said. "They have to get the processes right before they can get the enterprise AI right."