Ottawa has made faster AI adoption a national priority. Last month, the Prime Minister’s Office named a National Council on Artificial Intelligence to advise on the evolution of Canada’s AI strategy, including accelerating adoption, developing national champions and building sovereign AI infrastructure. The strategy also aims to give Canadians access to free AI literacy training – a $13-million National AI Literacy Initiative launched in September with courses for post-secondary students and K–12 educators.
Programs like that build capability over years. Employers need it sooner. Statistics Canada found that 44.4% of businesses using AI changed how they train or staff because of it, and many are competing for the same experienced talent. That raises a practical question: when to develop AI capability internally, and when to hire it.
When to train
So far, larger employers have been more likely to train than to hire. In the same Statistics Canada data, 68.1% of businesses with 100 or more employees that used AI in the past year gave existing employees AI training, and 51.7% did the same for executives. By comparison, 32.8% hired people with AI-related skills. Training tends to work best when:
- The need is broad fluency. Not every organization needs a dedicated AI executive. Often the priority is helping existing leaders understand where AI fits in the business and where it doesn’t, and training handles that well.
- Knowledge of the business matters most. Insiders who know the customers, processes and regulatory obligations can often apply AI faster than an outside hire still learning the business.
- There is a credible internal candidate. Someone who already has the business’s trust and some hands-on AI experience can often grow into the role faster than an outsider can build that credibility.
When to hire
Training existing executives builds shared fluency at the leadership table. It seldom substitutes for having done the work. The case for hiring is strongest when:
- The organization is about to scale. Moving AI from early experiments to something the business depends on, or growing an AI company past its first stage, often calls for someone who has managed that transition before.
- The role carries board-level or regulatory accountability. Understanding AI is different from managing its implementation, governance and risks. Where accountability is high, boards may benefit from an executive with a proven track record.
- A specialist team needs to be built quickly. PwC’s 2026 AI Jobs Barometer for Canada found postings for AI developer roles rose 74.4% in 2025. A leader who has built an AI team before knows which skills to hire for, how to structure the team and where to find the people, avoiding the false starts of a first attempt.
Doing both
Often the answer combines the two: one experienced AI leader, appointed from inside or outside, whose mandate includes building capability across the organization, supported by training for the wider team and executives.
Before looking outside, test internal candidates against the mandate: the scale of what they have delivered, the governance they have owned, and whether they can recruit the specialists the role depends on. Simulation-based assessment through our System-3 partnership can show how candidates handle realistic scenarios.
What it costs
Once the leadership mandate is defined, organizations need to consider what it will take to attract or retain someone capable of delivering it. Our Interim Report on Board and Executive Compensation in the AI Industry draws on fiscal 2024 compensation data from 116 publicly traded AI companies with market caps under US$2 billion. Median total pay for technology leaders was US$633,000, while pay at the 75th percentile reached US$1.48 million. Base salaries varied relatively little; equity did most to separate the highest-paid leaders. Hiring a technology leader out of an AI company means competing on long-term incentives as well as salary, and AI companies trying to keep those leaders face the same pressure.
Ottawa’s strategy will shape Canada’s AI workforce over the coming decade. Employers’ leadership choices will matter much sooner. Start with the mandate: what the role must deliver, own and build. Whether to train, hire or do both usually follows from that.
Bedford supports employers appointing senior AI leaders, from assessing internal candidates to running the search and setting compensation. If you’re deciding whether to train or hire, we’d welcome the conversation.