The Changing Canadian Workplace

Executive summary

Public discussion often treats labour shortages, AI, population aging and productivity as separate issues. This report examines them together because organizations experience them together: through the tasks people perform, the skills they need, the technologies they use, the schedules they work and the decisions managers make about how work is organized.

The report’s central interpretation is that these transitions converge through work design. National indicators describe the environment; organizations still have to decide how tasks are allocated, how workflows are structured, where technology belongs, what workers and managers need to learn, how experienced people are retained and what evidence should be used before a change is scaled.

This workplace lens also matters to the wider productivity debate. The OECD’s 2026 review of Canada’s productivity challenge emphasizes that demographic aging, digitalisation and artificial intelligence are structural shifts that require firms and workers to adapt. This report does not attempt to explain Canada’s productivity gap; it focuses on the organizational layer where those structural shifts become decisions about people, tasks, technology and work.

Five findings

1. Vacancy pressure has eased, but skill constraints remain visible.The broad vacancy environment is much less tight than in 2022, yet one-quarter of businesses still expect difficulty recruiting skilled employees. This distinction matters: aggregate labour-market easing does not eliminate occupation-, skill-, sector- or location-specific constraints.
2. AI adoption is accelerating, but there is no single Canadian AI transition.Business adoption, planned adoption and worker use are different measures. Adoption is concentrated in some sectors, low in others, and more than half of businesses still report no plans to use AI over the next year.
3. Technology acquisition should not be confused with productivity improvement.A 2026 Statistics Canada firm-level study found a large productivity gap between AI adopters and non-adopters in a benchmark model, but the gap narrowed substantially and became statistically insignificant after accounting for pre-existing productivity and complementary capabilities. The study does not establish a causal productivity effect from AI adoption.
4. Workforce aging raises questions about retention and the organization of work.More Canadians are working later in life, retirements are rising, and post-retirement work is often part time. The organizational implications include knowledge transfer, flexibility, training and the design of late-career work.
5. Work design is the bridge between national workforce trends and organizational performance.Recruitment, training, technology, schedules, workflows and decision rights interact. The practical challenge is to diagnose the specific constraint and adapt the work around it rather than defaulting to a single response.

Section 01

The workforce conversation has changed

The most useful starting point is precision. Canada is not experiencing the same labour-market conditions it experienced in 2022, but it would be equally misleading to conclude that workforce constraints have disappeared.

Job vacancies averaged 942,600 in 2022 and the vacancy rate reached 5.4%. In June 2026, job vacancies stood at 509,100 and the vacancy rate was 2.8%. There were 2.9 unemployed people for every vacancy in June 2026, compared with a ratio that fell as low as 1.0 in June 2022. Vacancy pressure has therefore eased substantially from the post-pandemic peak.

Skill pressure can persist inside a looser labour market

The Canadian Survey on Business Conditions tells a different but compatible story. In the third quarter of 2026, 25.2% of businesses expected recruiting skilled employees to be an obstacle over the next three months, 17.2% expected a shortage of labour force to be an obstacle and 15.5% expected retaining skilled employees to be an obstacle. These are business expectations, not measures of a national shortage, but they show that labour-related constraints remain material for many organizations.

The first-quarter 2026 skills analysis provides a complementary measure. Over the previous 12 months, 35.1% of businesses reported difficulty finding candidates with the skills needed to do the work. At the same time, 39.8% of businesses used in-house training to address skills gaps or deficiencies. Among businesses that had adopted new technologies over the previous three years, 85.7% provided training to support employees in adopting those technologies.

Source: Statistics Canada, Analysis on employee skills gaps, first quarter of 2026.

Sector context matters

The national averages conceal meaningful variation. In Q3 2026, recruiting skilled employees was expected to be an obstacle by 38.2% of manufacturing businesses, 33.6% of accommodation and food services businesses, and 31.7% of health care and social assistance businesses. A national workforce framework therefore needs to be diagnostic rather than prescriptive.

Source: Statistics Canada, Canadian Survey on Business Conditions, third quarter of 2026.

Section 02

AI is spreading quickly — and unevenly

Artificial intelligence is a prominent workplace technology story in 2026, but the data do not support a simple narrative of universal adoption.

In Q2 2026, 19.2% of businesses reported using AI to produce goods or deliver services over the 12 months preceding the survey, up from 12.2% in Q2 2025 and 6.1% in Q2 2024. In Q3 2026, 25.2% of businesses reported plans to use AI over the following 12 months, up from 14.5% a year earlier and 10.6% two years earlier.

Worker-level use is higher, but it measures something different. In March 2026, 35.9% of workers aged 15 to 69 reported using generative AI as part of their main job or business during the previous 12 months. A worker can use generative AI even when the organization does not report enterprise AI use, and a business can deploy AI systems that employees do not experience as generative AI.

Source: Use of generative artificial intelligence tools among Canadian workers, March 2026.

Sector differences are large

In Q2 2026, business AI use was highest in information and cultural industries (42.3%), finance and insurance (40.4%) and professional, scientific and technical services (32.4%), and much lower in agriculture (4.5%), wholesale trade (7.9%) and construction (9.2%). The implication is not that low-adoption sectors should imitate high-adoption sectors — the relevant question is whether a technology fits the actual tasks and operating model of a workplace.

Section 03

Technology adoption is not the same as productivity improvement

The most important productivity finding in the current Canadian AI evidence is cautionary: adopters look more productive, but the observed advantage becomes much smaller once the characteristics of adopting firms are taken into account.

A 2026 Statistics Canada firm-level study found that AI-adopting firms had labour productivity 16.8% higher than non-adopters in a benchmark specification controlling for standard firm characteristics. After the model added firms’ initial productivity, the estimated premium fell to 10.2%. After further controls for complementary capabilities — including research and development, cloud computing, data analytics, advanced robotics and ICT training — the association fell to 5.1% and was no longer statistically significant. The study also found no significant relationship between AI adoption and short-term productivity growth.

AI implementation creates organizational work

The Q2 2026 business survey shows what that integration can involve. Among businesses that had used AI in the previous 12 months, 44.4% reported making changes to training or staffing practices because of AI use. Larger AI-using businesses were substantially more likely to invest in training. These patterns do not prove which implementation model works best, but they make one point clear: adoption is not simply a software decision.

A better evaluation question

Instead of asking only whether an organization should ‘adopt AI,’ decision-makers can ask whether a defined workplace problem is suitable for a technology-enabled response and what would have to change around the tool for the response to work.

Section 04

An aging workforce changes the retention equation

Canada’s workforce is aging, but the workplace question is not simply how many people will retire. It is also how organizations retain experience, structure late-career work, transfer knowledge and support workers who remain economically active longer.

In 2025, the labour-force participation rate among Canadians aged 65 and older reached a record 15.2%, representing nearly 1.2 million seniors who were employed or looking for work. The average retirement age reached 65.4 years. At the same time, the 12-month average number of retirements rose from 183,900 in August 2012 to 276,800 in August 2025.

Retirement is not always a permanent exit from work

In 2023, 10.3% of Canadians aged 55 and older who had previously retired were working, up from 7.3% in 2019. Among those working after retirement, 73.3% were working part time and 32.8% were self-employed. These patterns should not be interpreted as evidence that older workers universally prefer part-time or self-employed work: financial circumstances, health, family responsibilities and job opportunities vary. They do show that the boundary between work and retirement is less binary for a meaningful group of older Canadians.

Source: Statistics Canada, Retirement and post-retirement employment among older Canadians, 2026.

Avoid simplistic claims about age and productivity

A 2026 Statistics Canada firm-level study found a non-linear association between the average age of a firm’s workforce and labour productivity. Depending on the model, firms with an average workforce age of 41 to 50 were 8.8% to 11.1% more productive than firms with an average age of 25 or younger, and 1.0% to 2.8% more productive than firms with an average age of 56 or older.

Those are firm-level associations, not estimates of individual worker productivity. They should not be used to rank workers by age. The more defensible lesson is that age composition, experience, skills, capital, technology and work organization interact. Retention strategy should therefore be connected to how work is structured and how knowledge is used.

Section 05

Work design is where structural change becomes operational

Labour-market conditions, AI diffusion, aging and productivity are macro-level stories. Organizations experience them through the design of everyday work — the allocation and sequence of tasks, staffing models, schedules, role boundaries, information flows, decision rights, training and technology.

Flexibility is one part of the picture — and it is unevenly available

Schedule control is one observable dimension of work design. For many operational jobs, workplace innovation therefore cannot primarily mean choosing when or where to work; the more relevant levers may be shift structure, staffing, task sequencing, equipment, workflow or the removal of administrative steps that do not add value.

Forum working definition

Workplace innovation is the evidence-informed redesign of tasks, workflows, roles, schedules, skills, technology and decision processes — ideally informed by the people doing the work — to improve how work is performed and experienced. Technology can be part of workplace innovation, but it is not a requirement and should not be treated as the objective in itself.

That definition deliberately links performance and people. A workplace change can increase output while also increasing burden, error, turnover or implementation complexity. Conversely, a change aimed at reducing burden can fail if it slows critical work or creates new coordination costs. Good workplace innovation therefore requires explicit trade-offs and measurement.

Section 06

There is no single Canadian workplace transition

A national report can identify common pressures without implying that every workplace is changing in the same way. The data already show pronounced differences by sector: business AI use in Q2 2026 ranged from 42.3% in information and cultural industries to 4.5% in agriculture. Worker generative-AI use in March 2026 ranged from 65.6% in professional, scientific and technical services to 16.3% in accommodation and food services. Recruiting skilled employees was expected to be an obstacle by 38.2% of manufacturing businesses in Q3 2026, compared with 25.2% across businesses nationally.

These differences matter because the same national policy or technology narrative can imply very different workplace decisions. A professional-services firm may need governance and workflow redesign around widespread generative-AI use. A health-care provider may be more constrained by staffing, schedules and task allocation. A warehouse may care more about process flow, physical demands and automation. A small service business may not see AI as relevant at all.

The role of a cross-sector framework is therefore not to prescribe one model. It is to help organizations distinguish the problem they share from the context that makes their response different.

Section 07

Six questions for workplace adaptation

The report’s evidence can be translated into six questions that move from a broad workforce concern to a specific, testable workplace decision.

1. What constraint are we actually trying to solve?.Separate headcount, skill, retention, workflow, technology, workplace-conditions and management-capability problems. They may overlap, but they are not interchangeable.
2. Where is scarce capacity being consumed?.Look for waiting, duplicate entry, avoidable hand-offs, rework, administrative burden, unclear decisions or tasks that consume skilled capacity without requiring skilled judgment.
3. Which tasks or workflows should change?.Focus below the job-title level, and involve the people closest to the work in diagnosing what actually happens. Technology and work redesign usually change specific tasks, information flows and sequences before they change whole occupations.
4. What capabilities are required before implementation?.Consider worker and manager training, data quality, governance, participation, technical integration, vendor support and the ability to interpret results.
5. What should be measured before scaling?.Establish a baseline and select a small number of measures tied to the problem: time, quality, output, burden, retention, exposure, error, user experience or another relevant outcome.
6. What did the intervention teach us?.End with a decision: scale, adapt, stop or investigate further. The value of a pilot is not that it occurred; it is the evidence it produces.

Section 08

What this means for Canadian organizations

The practical implication is not that every organization needs a workplace-innovation program. It is that workforce decisions become stronger when the problem, the work and the evidence are connected.

For employers

• Treat recruitment, skills development, retention and operational design as connected choices rather than separate functions.
• Evaluate technology at the task and workflow level. Ask what must change around the tool for value to be created.
• Build training and management capability into implementation plans instead of treating them as follow-on activities.
• Use bounded pilots where operational effects are uncertain, and establish the baseline before implementation.
• Connect late-career retention and knowledge-transfer strategies to the actual design of roles, schedules and tasks.

For industry and professional associations

• Use shared evidence to identify problems that recur across organizations rather than relying only on isolated anecdotes.
• Create structured forums for comparing implementation experience, including what failed or required adaptation.
• Where challenges are common, develop shared measurement or pilot frameworks that reduce duplication for individual organizations.

For workers’ compensation and prevention organizations

• Keep prevention connected to the broader design of work, alongside staffing, technology, workflow and administrative burden.
• Use workplace evidence to identify upstream conditions before adverse outcomes are visible, while maintaining a clear distinction between proxy measures, implementation activity and demonstrated outcomes.
• Support collaborative learning models that compare practical interventions without treating one technology as the default solution.

For governments and workforce institutions

• Do not treat aggregate vacancy easing as evidence that skill, retention or sector-specific constraints have disappeared.
• Support adaptability through skills development, worker transitions and diffusion capacity in addition to measures that affect labour supply.
• Improve evidence on what happens inside workplaces after technology adoption: training, task redesign, workflow change, implementation burden and outcomes.
• Use national frameworks to identify common issues while recognizing that implementation often requires sector- and place-specific evidence.

For research and innovation organizations

• Prioritize research that connects national workforce trends to task-, workflow- and workplace-level outcomes, where the current official evidence base is comparatively thin.
• Develop comparable measures of implementation, burden and outcomes so that organizations can learn across pilots and sectors rather than treating each implementation as an isolated case.
• Work with practitioners and workers to identify questions that are decision-relevant, measurable and transferable without assuming that one intervention will perform the same way in every workplace.

Section 09

Conclusion: from workforce pressure to workplace adaptation

Canada’s workforce debate is entering a more complex phase. The extraordinary vacancy pressure associated with 2022 has eased, but businesses continue to report skill and retention constraints. AI use is expanding rapidly in some parts of the economy while remaining limited or irrelevant in others. More Canadians are working later in life at the same time as the number of retirements rises.

These are not separate workplace stories. They affect who does the work, what skills are needed, which tasks are automated or augmented, how schedules and roles are structured, how experience is retained, and how organizations decide whether a change has actually improved performance.

That makes workplace adaptation a useful lens for the next stage of Canada’s workforce discussion. Adaptation does not mean constant redesign or technology adoption for its own sake. It means becoming more precise about the constraint, more deliberate about the response and more disciplined about the evidence used to judge results.

The most practical sequence is straightforward: define the workplace problem, establish a baseline, involve the people closest to the work, test a bounded response, measure what changed and decide whether to scale, adapt or stop. Applied consistently, that approach can help organizations turn broad workforce pressure into better workplace decisions.

Methodology, evidence standards and limitations

This report synthesizes publicly available Canadian labour-market, business, technology, retirement, productivity and working-conditions evidence available through September 1, 2026. Priority is given to Statistics Canada and other official public sources. OECD evidence is used to provide broader productivity context. The report is an evidence synthesis and practical interpretation; it is not a forecast, causal evaluation or sector-specific diagnostic.

Data-refresh note: the third-quarter 2026 Canadian Survey on Business Conditions was the final iteration of that survey. Future editions of the Forum’s indicators should therefore confirm successor Statistics Canada sources rather than assuming that the same quarterly series will continue.

Evidence principles

• Use the most recent published official data available at the report cut-off date and note the reference period of each measure.
• Do not treat measures with different populations, definitions or reference periods as directly comparable.
• Distinguish worker-level generative-AI use from business-level AI use and planned business adoption.
• Do not present descriptive survey results as proof of causality.
• Report firm-level productivity estimates with their model qualifications rather than as treatment effects.
• Do not interpret firm-level evidence on workforce age as individual worker productivity.
• Do not use national indicators to imply that every province, sector, occupation or workplace faces the same conditions.
• Identify practical implications as interpretation rather than statistical findings.

Key limitations

• AI definitions and use cases are evolving quickly, so adoption measures can become dated and may not capture informal use.
• Business surveys reflect self-reported experience and expectations rather than audited operational outcomes.
• National working-conditions data cannot describe task-level differences within every occupation or workplace.
• Sector averages can conceal meaningful differences by business size, region, occupation and establishment.
• The report does not attempt a complete causal account of Canada’s productivity performance.
• The report emphasizes cross-sector national evidence; provincial and sector-specific companion work would require separate analysis.