Commonwealth Bank sets a new bar for responsible AI at scale
Commonwealth Bank has released the first report by an Australian bank detailing how it ideates, builds, deploys and manages AI across the enterprise. The focus: responsible use, measurable outcomes and clear safeguards.
The report covers the past decade of AI progress at the bank and shows where AI is already at work-reducing scams and fraud, strengthening cyber security, improving customer experiences and detecting abuse in transaction descriptions. It's the kind of operating model managers ask for: structured, audited and tied to business results.
Why this matters for management
Industry estimates suggest AI could contribute $45-$115 billion annually to Australia's economy by 2030. Government policy continues to point business in the same direction, including the National AI Strategy 2030 from the Department of Industry, Science and Resources.
See Australia's National AI Strategy 2030
How CBA is doing it
Responsible AI, built into risk management
- Dedicated governance forum overseeing AI risk framework development.
- AI approach built on enterprise risk frameworks, guided by a Code of Conduct and six AI principles: environmental and social, fairness, transparency, privacy and data protection, reliability and security, and accountability.
- Toolkits and guardrails across the AI lifecycle with ongoing monitoring and annual reviews.
- Company-wide learning: an AI learning series launched in 2024, with 27,600+ employees engaged by end of 2025.
Fraud, scams and cyber: AI where risk is highest
- More than AU$900 million invested in FY2025 to combat fraud, scams, cyber threats and financial crime.
- AI models spot unusual patterns across over 20 million payments processed daily on average.
- 40,355 proactive warning alerts sent per day on average via the CommBank app.
- Result: customer fraud losses down over 20% in the first half of FY2026 vs the first half of FY2025.
- Partnership with Apate.ai uses AI-enabled bots to engage scammers in real time.
Customer experience and frontline enablement
- Customer Engagement Engine (in place since 2015) helps deliver more relevant experiences.
- Benefits finder has helped customers access 400+ rebates since 2019.
- Compass AI, a generative AI solution, helps frontline teams answer from the business bank knowledge base three times faster and has handled 500,000+ questions since July 2024.
- AI also flags abusive transaction descriptions to protect customers.
Building national capability through partnerships
- Founding member of the National Security Tech Alliance and founding partner in the Future Skills Organisation Skills Accelerator-AI.
- Collaboration with universities, including the University of Adelaide's Australian Institute for Machine Learning.
- Australian Institute for Machine Learning
- Investment in STEM and AI pathways for young Australians, with a focus on diversity and inclusion.
- Tech Hub launched in Seattle to connect with global technology leaders.
- In December 2025, the bank released an AI video series to help Australians build confidence with AI and introduced a technology skills program to help small businesses use AI to grow.
Leadership perspective
According to the bank's leadership, sharing practical lessons and risk controls is now part of the job. Stakeholders want visibility into how AI is used and how risks are managed. The message is consistent: build capability, encourage innovation and keep responsible adoption front and center.
Trust is a core theme. The bank links AI use to its risk foundations and principles, while continuing to collaborate with government, universities and industry. Expect ongoing refinement as the technology and threat environment shift.
What managers can apply this quarter
- Stand up a cross-functional AI governance forum with clear accountability and escalation paths.
- Adopt a concise set of AI principles and translate them into concrete controls: model checklists, approval gates, monitoring plans and annual reviews.
- Track outcomes that matter: alerts sent, fraud loss reduction, resolution time, model drift, audit findings and customer impact.
- Invest in skills at scale. Blend foundational AI literacy with role-based training for risk, data, engineering and frontline teams.
- Target high-leverage use cases first: fraud/scam detection, cyber threat detection, customer support acceleration and content retrieval for staff.
- Partner where it makes sense (universities, startups, vendors) and run structured pilots with clear success criteria.
- Measure trust. Run longitudinal research or consistent surveys to track customer and employee confidence in AI-enabled services.
Key numbers to brief your board
- AU$900m+ invested in FY2025 for fraud, scams, cyber and financial crime protection.
- 20m+ payments processed daily on average; 40,355 proactive alerts sent per day on average.
- Fraud losses down 20%+ (H1 FY2026 vs H1 FY2025).
- 27,600+ employees engaged in AI learning by end of 2025.
- 500,000+ frontline questions answered by Compass AI since July 2024.
Research and transparency
This week, the bank also launched a multi-year research initiative with Melbourne Business School to track how Australians perceive, use and trust AI in banking over time. The full "Our Approach to Adopting AI" report is available on the bank's website.
Next step for your team
If you're building capability and need structured learning paths by role, see curated AI programs for managers and teams:
AI courses by job role - Complete AI Training
Your membership also unlocks:
