AI is increasingly being used across Canadian factories, construction sites, farms and other heavy-industry workplaces, but employer support has not kept pace with adoption. A new report from Punchcard Systems and Angus Reid finds that nearly half of heavy-industry workers already use AI on the job, while more than two-thirds have received no formal training or guidance on how to use it safely and effectively.

Artificial intelligence is moving deeper into Canada’s frontline workforce, but employers are struggling to provide workers with the training and governance needed to use it effectively, according to new research from digital innovation consultancy Punchcard Systems and Angus Reid.

The State of AI at Work in Canada Report surveyed 1,100 working Canadians across industries to examine how AI is being used, which tasks employees trust it to perform, and how organizations are supporting workers as AI becomes part of everyday jobs.

The findings highlight a growing gap between AI adoption and workforce readiness.

Nearly half, or 49%, of Canadian heavy-industry workers said they already use AI for work. Yet 68% said their employers had provided no formal AI support or training. Construction workers reported the largest gap, with 87% saying they had received no formal support.

Agriculture followed at 77%, while 67% of manufacturing workers and 65% of transportation and logistics workers reported receiving no formal AI support.

The gap matters because AI is already changing how people perform their jobs. Across Canada, 44% of workers said AI has affected their role. That figure rises to 54% among transportation and logistics workers, 52% in professional services, 45% in manufacturing and 45% in oil and gas.

AI is producing measurable productivity gains for some workers. Those who save time with AI reported gaining an average of 6.1 hours per week.

But the productivity calculation is more complicated than simply measuring time saved.

Ninety-four percent of Canadian workers who use AI said they rework its output before using it. Nearly one in five, or 19%, said reviewing or fact-checking AI-generated work has become a regular part of their job.

The technology can also introduce operational risk. Eighteen percent of workers said AI-generated output had caused a problem at work, while another 24% said they had caught AI errors before those errors caused a problem.

For employers, that creates a workforce-management challenge: AI can reduce time spent on routine work while simultaneously creating new responsibilities around verification, quality control and accountability.

The research also points to a significant governance gap. Only 6% of workers who use AI said they have been given formal responsibility for AI governance, policy or safe use within their organization.

Workers whose employers provide formal AI support were three times as likely to have that responsibility, at 9%, compared with 3% among workers whose employers provided no formal support.

That relationship suggests AI governance is not simply a technology-policy problem. It is also an organizational-design and workforce-development issue.

Heavy-industry workers appear to recognize the distinction between low-risk assistance and high-stakes decision-making. Seventy-seven percent said they trust AI to look up routine procedures, product specifications or how-to information, while 71% trust it to help write reports, logs or other job paperwork.

Safety decisions are different.

Nearly 45% of heavy-industry workers said they do not trust AI at all with decisions affecting worker safety. Even among construction workers, who showed comparatively high willingness to use AI for technical tasks, none said they would completely trust AI with safety decisions.

At the same time, 64% of construction workers said they trust AI to read technical drawings when they can review its output, while 63% expressed similar trust for interpreting Canadian codes and standards.

The distinction illustrates an emerging model for workplace AI: human-in-the-loop automation, in which AI handles information-intensive or repetitive tasks while employees retain responsibility for decisions with safety, financial or operational consequences.

That approach may become increasingly important as organizations deploy AI outside conventional office environments. Factory workers, construction crews, agricultural employees and logistics teams operate in environments where incorrect AI output can have consequences beyond an inaccurate document or inefficient workflow.

The Punchcard research also suggests employers need to rethink AI implementation at the process level rather than treating AI as another software deployment.

Adding a chatbot or generative AI assistant to an existing workflow may produce limited gains if employees do not know when to use it, how to verify its output or who is accountable when it fails.

For HR and workforce leaders, the challenge is therefore broader than AI training. Organizations need clear policies, role-specific guidance, governance responsibilities and workflows that define where automated systems can act and where employees must make the final decision.

That becomes particularly important as workers increasingly encounter AI without formal organizational support.

The Canadian findings suggest the next phase of workplace AI adoption will be determined less by whether employees have access to the technology and more by whether organizations can redesign work around it responsibly.

For HR technology vendors, that creates demand for tools that combine AI enablement with skills development, policy management, workforce analytics and governance. For employers, it means treating frontline employees as participants in AI transformation rather than simply users of newly deployed software.

The productivity opportunity is significant, but the report’s findings indicate that organizations will need to invest in the people and processes surrounding AI if those gains are to translate into sustainable workforce improvements.

Market Landscape

Workplace AI is moving from experimentation toward operational deployment, but frontline industries face different requirements from knowledge-work environments. Manufacturing, construction, transportation and agriculture require role-specific training, stronger human oversight and clearer boundaries around safety-critical decisions.

The emerging HRTech opportunity extends beyond AI assistants into AI workforce enablement: training employees, establishing governance responsibilities, monitoring adoption, redesigning workflows and measuring whether automation actually improves productivity.

The research also highlights a potential divide between organizations that simply provide AI tools and those that integrate AI into formal workforce strategies. Employers that connect technology deployment with training and accountability may be better positioned to capture productivity gains without shifting excessive verification or risk-management work onto employees.

Top Insights

• Nearly half of Canadian heavy-industry workers use AI, yet 68% report receiving no formal employer support or training.
• Construction has the largest training gap, with 87% of workers reporting no formal AI support from their employers.
• Workers who save time with AI recover an average of 6.1 hours per week, but 94% still rework AI-generated output.
• Only 6% of AI-using workers have formal responsibility for AI governance, policy or safe use at their organizations.
• Workers generally trust AI for routine information and documentation more than safety-critical decisions requiring human judgment.

Join thousands of HR leaders who rely on HRTechEdge for the latest in workforce technology, AI-driven HR solutions, and strategic insights