Canadian firms race to adopt AI despite readiness gaps

Canadian companies are rapidly adopting agentic AI, even as most say they are not fully prepared to deploy and govern it. Research from SAP and Oxford Economics found that 66 per cent of organisations in Canada have already piloted the technology.

The findings highlight a sharp gap between investment plans and operational readiness as businesses move beyond generative AI tools toward software agents that can take actions inside company systems. Still, 97 per cent of Canadian organisations do not consider themselves fully ready for the next phase of AI deployment.

That push is backed by sizeable spending expectations. The average Canadian organisation expects to spend CAD $39.4 million on AI this year and forecasts a 20 per cent return on that investment. Within two years, expected returns are projected to rise to 38 per cent, amounting to an additional CAD $20.2 million in value per company.

Yet those expectations are running into practical problems. Among Canadian organisations already using agentic AI, 64 per cent said integration was harder than expected, while 54 per cent reported reliability or quality issues as usage increased. Almost half, 48 per cent, said agents had taken incorrect actions, and 44 per cent reported inconsistent outcomes for the same process.

Control gaps

The figures suggest many businesses are rolling out autonomous AI systems before putting basic oversight in place. Only 39 per cent said they maintain an agent registry, and just 28 per cent have cost or usage controls. Nearly half, 46 per cent, said they do not have a human-in-the-loop process for autonomous workflows.

Seventy per cent of Canadian business leaders said they either agree or are unsure that they are deploying AI agents faster than governance can keep pace. That matters because agentic AI differs from earlier AI tools in one key respect: it does not just produce information, but can also make decisions and carry out tasks on behalf of users.

SAP argues the issue is less about the underlying technology and more about whether businesses have the data, processes, and accountability structures needed to use it safely. The research found that many companies still lack those basics despite confidence in their preparedness.

While 65 per cent of respondents said they believe they are data-ready, 73 per cent also said they struggle with incomplete or inconsistent data. Low-quality outputs are already affecting operations, with 71 per cent of organisations reporting rework and delays linked to weak AI results.

Strategy questions

The report also points to fragmented decision-making. Only 18 per cent of Canadian companies said they take a centralised, strategic approach to AI. By contrast, 39 per cent use department-led approaches and 35 per cent are pursuing piecemeal efforts.

That fragmentation can make it harder to connect AI initiatives to broader business priorities, especially as autonomous agents are introduced across multiple functions. The research suggests many organisations still treat AI as a collection of separate tools rather than as part of core operating systems.

Skills are another pressure point. Seventy-eight per cent of business leaders said they are not convinced their upskilling efforts are keeping pace with the speed of AI change. More than half of organisations, 56 per cent, said unauthorised use of AI tools occurs at least occasionally, indicating that internal controls are struggling to match employee demand for new systems.

Cathy Tough, Country Manager, SAP Canada, said the challenge is preparing organisations to derive useful results from the technology rather than simply deploying it.

"The technology is ready. The challenge is getting the business ready for the technology. Getting measurable value from AI requires a different approach," said Tough.

"AI that lacks business context creates activity without outcomes, and at worst introduces risk. Businesses in Canada, large and small, need to connect AI directly to the core data and processes that run their organizations and make sure it has the context and governance to drive trusted results. That's what we call the Autonomous Enterprise, where agents, processes, and people work as one, backed by trusted governance, to achieve real outcomes."

Governance focus

Governance was one of the weakest areas of preparedness. Fewer than one in five respondents said their internal skills are fully ready for AI governance, at 17 per cent, while only 16 per cent said their processes are fully ready.

Anthony Robinson, Chief Technology Officer, SAP Canada, said companies need more structure around how autonomous systems are deployed inside business operations.

"The path to value isn't just more AI, it's smarter AI operating within a trusted framework. SAP's model for the Autonomous Enterprise is simple: humans direct, assistants coordinate, and agents execute," said Robinson.

"This ensures that humans set the strategy while AI handles the orchestration and execution, all within a governed, auditable environment with humans in the loop at key moments. By embedding AI into core business processes, rather than bolting it onto a shaky foundation, we close the exact readiness gaps this study highlights and give businesses the confidence to scale."

The global research surveyed 2,600 business leaders across 13 countries, including 200 in Canada, spanning midmarket organisations and larger enterprises.