A Matter of Trust: Why CPAs Have a Critical Role in AI Governance
September 11, 2026
For finance leaders in the public sector, AI presents more than just the opportunity to automate work. It is to determine where AI can generate value while ensuring the systems behind it remain secure, explainable and accountable.
Émile Wandji, CPA, MBA, Director General of Finance and Administration & Deputy CFO at the Public Service Commission of Canada (PSC), and his team identified a bottleneck in government travel. Employees regularly turned to finance for help interpreting the National Joint Council Travel Directive, a set of parameters that governs the reimbursement of travelling expenses for public service employees, which then led to hundreds of repetitive questions about per diems, documentation, approvals and other policy requirements.
The impact extended well beyond an overflowing inbox. Delayed responses could lead to later travel bookings and higher costs while trained financial professionals spent valuable time answering routine questions rather than concentrating on higher order strategic work, including risk management and controllership.
The PSC developed a bilingual AI agent connected directly to government travel policies. The result was a 90 per cent reduction in routine emails and more time for employees to focus on the work that required their strategic insight.
It is a compelling example of AI delivering measurable operational value. But for Wandji, implementing the technology is only part of the equation.
Innovation requires governance
Public sector AI operates within a unique environment. Governments provide services citizens can’t obtain elsewhere, placing public trust at the centre of how AI systems must be governed. The public sector also has access to huge amounts of sensitive data, so ensuring that cybersecurity is tight is also a top concern.
"In the private sector, questions of success typically revolve around how fast something can create shareholder value," explains Emile. “For the public sector, it goes beyond efficiency: Is the system fair? Is it explainable? Is it lawful? Does it treat citizens equally?”
At the PSC, safeguards begin before an AI tool is deployed. Potential risks are assessed upfront, privacy reviews examine data flows, personal information is removed where appropriate and AI agents operate within secure government cloud environments. Role-based access controls also ensure employees cannot use an AI tool to access information they would not otherwise be authorized to see.
Even with these guardrails, Emile stresses that human oversight must remain fundamental: “Automated systems can assist with analysis and recommendations, but a qualified officer must remain responsible for critical decisions.”
Travel is just one example of how the public sector is deploying AI. The PSC has developed agentic AI to assist program managers with procurement, including generating an initial statement of work that can then be reviewed and refined by a person, reducing the process time to approximately one minute.
Why CPAs belong at the AI governance table
For Emile, AI governance is a natural area for CPAs to lead.
AI may once have been viewed primarily as an IT or data science initiative. However, its most important organizational questions involve internal controls, data integrity and risk management.
"Just like CPAs validate financial statements, they can validate AI systems,” argues Emile.
CPAs can help design the checks and balances surrounding AI-enabled financial processes, independently evaluate the quality and reliability of outputs and determine whether investments are producing measurable returns. They can also bridge the gap between technical teams developing AI solutions and executives responsible for organizational strategy.
"CPAs do not need to become machine-learning engineers,” says Emile. “But they do need sufficient data and model literacy to question assumptions, assess the reliability of source data and recognize the limitations of AI-generated analysis."
As routine work becomes increasingly automated, professional value will shift toward judgement, interpretation and strategic leadership.
This lesson extends well beyond government. Organizations adopting AI need governance from the beginning, clear traceability and controls that treat fairness, transparency and privacy as operational requirements rather than afterthoughts.
For CPAs, that creates an opportunity not simply to adapt to AI but to help determine how organizations use it responsibly.
Émile Wandji will share insights from the Public Service Commission of Canada’s AI transformation at CPA Ontario’s The Edge Conference: The Future of Finance Leadership on September 29, 2026.