A financial services organisation wanted to increase its use of AI. Its established systems were isolated, and staff were already using publicly available AI services. Building business-specific, context-aware AI agents and reusing existing RPA investments gave the organisation a way to accelerate while maintaining compliance.
Challenge
Established system constraints drive shadow AI use, which increases risk in a regulated environment.
- Core services depend on multiple, costly-to-integrate established systems.
- Fragmented data across platforms limits reuse.
- Leadership demand is accelerating while staff already use public AI tools.
Guardrails
Utilise existing technology investments while strengthening controls.
- Integrate around, not replace multi-year technology investments.
- Apply strict controls to personal data while accommodating evolving AI initiatives.
Approach
A governed AI layer to connect data, AI agents and core systems.
- Unify governed data access through an open data lake and common cleansing.
- Utilise RPA to execute transactions in established systems.
- Centralise agents on one platform with approved models and controls.
Impact
Scale AI use safely while retaining overall enterprise control.
- Establish an AI centre of excellence that responds to business demand.
- Standardise security, risk and third-party approval across agents.
- Centrally manage model choice, usage and cost.