Scaling AI Fluency Across a Large Enterprise Workforce

Commonwealth Bank of Australia (CBA) set out to embed conversational AI in the daily work of nearly 50,000 staff. The initiative required a clear governance model, easy‑to‑use integrations, continuous learning, and measurable outcomes for customers.

Technical Solution

The rollout combined a centralized policy hub, pre‑built connectors to core business systems, a tiered training curriculum, and pilot projects that demonstrated value in high‑impact moments such as fraud response and customer service.

Governance and Security Framework

CBA established an enterprise‑wide policy that mandates encryption at rest and role‑based access control. The framework draws on guidance from the OpenAI GPT‑4 System Card and aligns with the cloud computing architecture standards used across the bank.

Connector Ecosystem

Pre‑configured connectors link ChatGPT Enterprise to CRM, ERP and fraud detection platforms, reducing the need for custom code. Teams can add new data sources through a low‑code portal, following best practices outlined in Choosing the Right AI Model for Your Project.

Training and Community Programs

A blended learning path starts with mandatory security briefings, then moves to hands‑on labs and peer‑run forums. Daily micro‑tasks encourage users to experiment, while leadership models the behavior by publishing AI‑enhanced briefs.

Agent‑Powered Use Cases

Early pilots focused on customer service bots that surface relevant account details, and fraud‑alert agents that triage suspicious activity. Results are tracked against key performance indicators published in the bank’s internal Domain Authority dashboard.

Monitoring and Continuous Improvement

Telemetry feeds into a centralized dashboard that flags anomalies such as hallucinations, referencing the AI hallucination problem. Feedback loops inform prompt‑engineering refinements documented in the AI Prompt Engineering guide.