The productivity argument for AI tools usually centres on speed: same work, less time. Christine Andrew, KPMG Canada's managing director for AI enablement, is making a different case.
Andrew has been using Microsoft 365 Copilot longer than most. She led an early-access pilot of 40 licences nearly three years ago, helped expand the trial to 1,000 users, and was part of the firmwide rollout to all 12,000 KPMG Canada employees in 2025. Her conclusion is not that the tool made her faster. It is that it changed what she spends her time on.
"There is a productivity play there, but I think the real advantage is being able to reallocate the time to more strategic activity and higher value activity," Andrew said in an announcement from KPMG Canada.
What she actually uses it for
The practical applications Andrew describes are unglamorous but telling. She uses Copilot in Teams for meeting recaps and catch-up prompts when her calendar runs double or triple-booked. She uses Copilot Chat as a thought challenger, a tool to stress-test strategies rather than simply execute them. She credits its research features with strengthening a recent keynote enough to earn an invitation to return.
Each of these use cases compresses a specific type of operational drag: the time spent getting up to speed, the time spent preparing rather than thinking, the time spent on logistics rather than judgment. What remains, Andrew argues, is more advising, more reflection and more time on harder client problems.
The deployment story
The scale of KPMG Canada's rollout gives Andrew's account more weight than a typical vendor case study. A firmwide deployment to 12,000 employees is not a pilot. It is a commitment, and one that carries real risk for a firm whose reputation rests on ethics and risk standards.
Andrew says Copilot's integration with Microsoft's ecosystem made responsible use easier to enforce, keeping the tool aligned with KPMG's internal standards rather than creating a new compliance problem alongside the productivity gains.
The argument worth taking seriously
Related reading
- Microsoft warns AI agents risk becoming "double agents" as it unveils security controls at RSAC
- Meta hires founders of AI startup Dreamer in latest push to build autonomous agent technology
- OpenAI builds real-time monitor to catch misaligned behaviour in its own AI coding agents
Most AI adoption conversations focus on the wrong metric. Time saved is easy to measure and easy to sell. What is harder to quantify is whether the time freed up gets spent on anything more valuable, or simply absorbed by the next item on the list.
Andrew's account suggests the more interesting question is not how much faster AI makes knowledge workers, but whether it changes the quality of the work they do when the routine tasks are handled. Her answer, three years in, is yes. The work is different, not just faster.
The recap
- Christine Andrew says Copilot frees time for higher-value work
- Pilot began with 40 licenses, expanded to 1,000 users
- Copilot deployed to all 12,000 employees across the firm in 2025
