Swiss Pilot Cuts Client Onboarding From Months to Days With AI Agents
Published on 09/04/2026 at 04:34 | Editorial boerse-global.de
A collaboration between a Swiss private bank, a global IT services firm, and a major cloud provider has demonstrated that artificial intelligence can slash one of banking's most notoriously slow processes down to a fraction of its former duration.
The Incore Bank, Kyndryl, and Google Cloud announced on September 2, 2026, that their joint proof of concept had successfully tested agent-based AI for know-your-customer (KYC) procedures — the identity verification and risk screening that financial institutions must complete before taking on new clients.
Gemini Models Handle Risk Screening
At the heart of the trial sat Google's Gemini models, which were deployed to run automated risk assessments within the KYC workflow. The technology functioned as an agent-based system capable of independently parsing complex data structures and preparing them for downstream processing.
Each partner brought a distinct piece to the puzzle. The Incore Bank defined the regulatory framework and domain expertise, Kyndryl handled system integration, and Google Cloud supplied both the infrastructure and the AI models themselves. Their shared objective was reconciling the strict compliance demands placed on financial institutions with what modern technology can actually deliver.
Near-Perfect Data Extraction
One of the standout findings concerns accuracy. According to the companies involved, the proof of concept achieved up to 99 percent precision when extracting relevant information from submitted documents. That level of reliability matters enormously in highly regulated environments, where flawed data extraction can trigger serious legal and financial fallout.
The traditional approach — manual or partially automated processing — has long been recognised as both error-prone and slow. By capturing customer data precisely and cross-referencing it automatically against risk profiles, the system reduces the potential for human mistakes and improves data quality for subsequent verification steps.
Efficiency Gains Reshape Onboarding
Speed was the other major focus of the pilot. The results indicate that AI-driven automation can compress the entire onboarding timeline from several months down to just a few days.
That acceleration carries direct operational consequences. Protracted review procedures frequently stand in the way of establishing new business relationships. Shorter waiting periods improve scalability for the bank and free compliance teams from repetitive administrative drudgery.
Broader Lessons for Compliance Work
The project's success points to wider applications for combining cloud infrastructure with specialised AI agents in administrative compliance tasks. Automating due-diligence checks illustrates how technological support can translate complex legal requirements into more efficient workflows.
While the trial was rooted in the banking sector, its outcomes reinforce a broader trend toward digitising background checks and identity verification across industries. For organisations operating under stringent regulatory regimes, integrating such AI models offers a path to maintaining security standards while dramatically accelerating process speed. The experience gathered by the Incore Bank, Kyndryl, and Google Cloud now provides a solid evidence base for similar deployments in other demanding fields.
