Wanted: Proper AI training

Employers and HR leaders had better come up with the appropriate artificial intelligence (AI) training for workers soon.

That’s because most Canadian workers say they want structured, employer-led training to use artificial intelligence (AI) on the job, not informal trial and error. Currently, 79% of Canadian job seekers say companies need to formally train employees on how to use AI, rather than expecting them to learn on their own, report Express Employment Professionals and The Harris Poll.

Hiring managers largely agree, with 77% saying formal AI training should be a company priority.

When asked what that training should look like, hiring managers emphasize practical, workplace-based learning. The most frequently cited approach is on‑the‑job training focused on working alongside AI, identified by 38% of hiring managers as the best way to build capability in real work contexts. Another 38% point to dedicated training for skills AI cannot replace, highlighting the continued importance of human capabilities such as critical thinking, relationship management and ethical judgement.

Nearly one‑quarter of hiring managers (24%) highlight apprenticeship and internship programmes that include AI training as an important way to prepare future talent.

Workers say that AI is “adding pressure to have even higher productivity,” according to a previous report.

AI adoption in Canada

The call for formal training comes as nearly two‑thirds of Canadian companies (63%) now use AI, with 19% reporting regular use, based on the Express report based on two surveys: one  among 504 Canadian hiring decision-makers conducted Nov. 3 to 19, 2025, and the other among 502 adults ages 18 and older conducted Nov. 7 to 21, 2025.

More than half of employed Canadian job seekers (53%) say their company uses AI at work, including 41% who say they personally use it at least sometimes. Among companies already using AI, reliance is intensifying: a large majority (73%) say their dependence on AI tools has significantly increased over the past year.

Yet policy and guidance have not fully kept pace. According to Express Employment Professionals, 58% of hiring managers say their company has policies regulating AI use, and 73% of employed job seekers say their employer has some form of AI policy. However, only 29% of companies provide a list of approved or preferred AI tools, while 37% allow employees to use any AI tools they are familiar with and 18% report a mix of approved and open‑use tools.

“AI adoption is moving faster than most organizational change ever has,” says Bob Funk Jr., CEO, President and Chairman of Express Employment International. “What this data shows is that companies have focused on getting the technology in place, but not enough on helping people use it effectively. Training is what determines whether AI becomes a source of real productivity or just another tool employees are left to navigate on their own.”

Despite gaps in formal training, workers are generally optimistic about AI’s role in building skills and careers. More than half of job seekers (57%) say their company’s AI tools can help bridge skills gaps. In addition, 67% say they are likely to seek additional training in response to AI advancements.

On the employer side, many hiring managers believe the technical base for AI‑related learning is already in place: 59% say their company has the tools needed to train new hires in AI‑driven workflows.

AI adoption is key to job security, according to a previous report.

How can companies succeed in upskilling employees on AI?

Here’s how employers can effectively train workers on AI, according to Paul Thomas, founder and consultant at The Human Co:

• Replace one-off workshops with ongoing “AI clinics”. Offer regular drop‑in sessions or office hours where employees bring real tasks and problems, get help using AI on the spot, and know they can return as their work evolves, instead of relying on single “learning days” or bootcamps that are forgotten within weeks.
• Train around workflows, not generic prompting skills. Start with the “grumbles” in each role (backlogs, hated tasks, bottlenecks), then show employees how to use AI to fix those specific workflow pain points, revealing the real skills gaps (e.g., accuracy checking, compliance review) far better than generic prompt-writing courses.
• Build psychological safety by giving employees agency. Address fears about job loss directly, teach where AI fails, and position staff as the essential quality‑control layer who must review, question and correct AI outputs, while mapping clear future roles and human skills that AI cannot replicate.
• Co-create reskilling paths with at‑risk employees. For roles likely to be automated, work transparently with affected employees to define what their job looks like “on the other side” of AI—identifying concrete new skills and responsibilities rather than offering vague reassurances decided in isolation by HR.
• Put enabling governance in place instead of bans. Accept that “shadow AI” is already happening and replace blanket restrictions with approved tools and clear guardrails so staff can experiment safely, rather than driving sensitive work into unmanaged consumer apps on personal devices.
• Create a cross-functional AI council to steer training and use. Bring Legal, IT and business units together to approve use cases, set standards and provide “safe harbour” for experimentation, shifting AI from an IT project to a business-wide capability programme that supports responsible, scalable employee training.

Admin workload is preventing HR from contributing more strategically to business success - and agentic AI can help, according to a previous study.