Incore Bank, Kyndryl and Google Cloud Report KYC Onboarding Proof of Concept
Demonstrates a concrete agentic-AI governance architecture for high-consequence automated decisions - useful reference for APS teams designing similar workflows.
Key points
- Incore Bank, Kyndryl, and Google Cloud completed an agentic-AI proof of concept for KYC onboarding and risk assessment.
- The PoC architecture combines document extraction, policy-as-code controls, explainable risk scores, and auditable decision records with human oversight.
- This is a private-sector banking PoC with no direct Australian regulatory or APS governance angle at this stage.
Implications for Australian agencies
- Monitor APS teams working on automated decision-making or identity verification workflows may want to monitor how this agentic architecture - combining policy-as-code, audit trails, and human oversight - matures toward production deployment.
- Consider Agencies designing ADM governance frameworks could consider this PoC as a reference example of how explainability, auditability, and human-in-the-loop controls can be embedded in agentic workflows from the outset.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"Incore Bank, Kyndryl and Google Cloud Report KYC Onboarding Proof of Concept"
Source: Let's Data Science – AI Governance
Published: 2 September 2026
URL: https://letsdatascience.com/news/incore-bank-kyndryl-and-google-cloud-report-kyc-onboarding-p-89e56510
Incore Bank, Kyndryl, and Google Cloud have announced the completion of a proof of concept using Google Gemini models with Kyndryl's Agentic AI Framework to automate KYC customer onboarding and risk assessment in banking. The PoC reported up to 99% accuracy in extracting data from onboarding documents and claims potential to reduce onboarding timelines from months to days. Crucially, the design retains human oversight for judgment-intensive cases and incorporates policy-as-code guardrails, audit trails, and explainable risk scoring. These figures describe a prototype only and have not been independently validated or deployed in production.
Implications for Australian agencies:
- [Monitor] APS teams working on automated decision-making or identity verification workflows may want to monitor how this agentic architecture - combining policy-as-code, audit trails, and human oversight - matures toward production deployment.
- [Consider] Agencies designing ADM governance frameworks could consider this PoC as a reference example of how explainability, auditability, and human-in-the-loop controls can be embedded in agentic workflows from the outset.
Retrieved from SIMS, 15 November 2026.