AI agents help banks assess regulatory changes

A recent webinar hosted by Finextra with Workday examined how AI agents can help financial institutions link contracts, supplier records, financial data and internal policies so specialists can assess regulatory changes faster and coordinate responses across the business.

Panelists on the session included Patrick O’Donoghue of Workday and moderator Jane Cooper of Finextra. Panelists noted regulatory updates often trigger work across finance, procurement, legal, HR, operations and IT, and that the main delay is the time required to find and connect information held in multiple systems.

Panelists described how AI agents can gather relevant business context by extracting contract clauses from document repositories, joining supplier details from procurement systems, mapping reporting fields in finance platforms and pulling organisational structures from HR systems. The agents then present a narrowed set of findings for legal, compliance and finance specialists to review.

Examples of agent tasks discussed included identifying contracts with non-compliant terms, flagging customers or suppliers affected by sanctions, listing reporting metrics that need revision, and estimating potential cost or capacity impacts for operational teams. Panelists emphasised that agents are support tools and that humans remain responsible for interpretation and final approvals.

Technical integration steps outlined were connecting agents to contract management, ERP and HR systems, establishing secure data feeds and building audit trails for automated findings. Operational steps included defining ownership of follow-up actions, creating escalation paths for identified issues and training subject-matter experts to validate agent outputs.

The panel identified compliance and legal, finance and procurement, HR and operations, IT and data governance, and senior risk and regulatory affairs as key stakeholders. Successful projects begin by mapping which groups own which data sources and decisions and by prioritising use cases that produce measurable time savings or reduce compliance risk.

An end-to-end example presented by panelists began with automated monitoring for regulatory updates. After a change is detected, an agent performs an impact scan across contracts, supplier records and reporting schemas and produces a ranked list of affected items. Subject-matter experts review the findings, decide on contractual or process changes, assign tasks to operational teams and use the agent to track remediation and collect evidence for audits.

Panelists noted implementation limits including data silos, poor metadata, upfront integration and data-cleaning work, the need for explainability and traceability for regulators, and the requirement for strong security and privacy controls when agents access sensitive records. Panelists added that institutions often pilot agents on narrowly defined regulatory tasks before wider rollout.