Incore Bank has completed a proof of concept that applies agentic AI to customer onboarding.
The Zurich-based bank worked with Kyndryl and Google Cloud to test whether AI agents built on Google’s Gemini models could speed up Know Your Customer (KYC) checks.
The goal was to keep risk decisions explainable and under human oversight.
The proof of concept applied Kyndryl’s Agentic AI Framework together with Gemini.
It added a semantic layer built on Kyndryl’s policy-as-code capability, supported by guardrail agents.
The setup supports end-to-end risk assessments, covering both onboarding and ongoing reviews.
It also monitors for changes in a customer’s circumstances over time.
Several AI agents work across structured and unstructured data, internal banking systems and external sources.
According to the companies, the agents extract and validate customer information, gather evidence and identify risk factors.
They also generate explainable risk scores and create auditable decision records for compliance teams.
The proof of concept achieved up to 99% accuracy in the automated extraction of data from onboarding documents.
The companies said this showed potential to cut onboarding times from months to days.
It also freed compliance staff to focus on judgement-intensive cases while keeping transparency and human oversight in place.
“The proof of concept we completed with Kyndryl and Google Cloud shows how cloud-enabled agentic AI can help banks assess both prospective and existing customers more efficiently while maintaining the rigorous oversight and controls required in a regulated environment,”
said Mark Dambacher, CEO of Incore Bank.
“Agentic AI is the next stage in the evolution of enterprise AI, orchestrating complex business workflows. In highly regulated industries such as banking, success depends on strong governance, explainability, auditability, secure data access and human oversight,”
said Jacqueline Wild, Vice President and General Manager of Kyndryl Alps, the division covering Switzerland and Austria.
Featured image credit: Edited by Fintech News Switzerland, based on image by Who is Danny via Magnific
