Expert says issue is not a technology problem: 'It is a work design and people strategy problem'
Most artificial intelligence (AI) projects fail, and it’s mostly due to a lack of employee engagement—not technical issues—says one expert.
Overall, 80 per cent of AI projects fail at the human level, according to data provided by Work Time Reduction.
That’s because organisations “overlook the importance of employee buy-in and structured support,” says Joe O’Connor, CEO.
Citing the EU’s Generative AI Outlook Report, he notes that while 85 per cent of enterprises view AI as a strategic priority, fewer than 20 per cent have redesigned work to support its adoption.
AI: 'people strategy problem'
Only 11 per cent of employees report receiving structured support to use AI tools.
“This gap between ambition and execution is not a technology problem—it is a work design and people strategy problem,” O’Connor says.
He also notes that the main barrier to AI adoption is time, not cost or skills. Organisations that provide protected time for AI learning and experimentation are five times more likely to report successful adoption. However, a LinkedIn study found that 51 per cent of professionals feel that keeping up with AI is like taking on a second job.
“You cannot build AI capability without first creating capacity,” says O'Connor. “Asking people to accelerate AI adoption off the side of an overflowing desk is like driving a Formula 1 car in rush-hour traffic.”
He also references an Abacus survey, which found that only 17 per cent of Canadians believe that “unleashing the full potential of AI” would benefit them personally, while 31 per cent believe it would be detrimental and the rest are uncertain or indifferent.
Canada’s workforce trails much of the world in both confidence and literacy related to AI systems, according to a previous KPMG report.
AI sabotage, shadow AI on the rise
O’Connor also highlights two emerging trends in the workplace: AI “sabotage” and “shadow” AI. He notes that research from Writer shows that 31 per cent of employees—rising to 41 per cent among millennials and gen Z—are actively sabotaging their company’s AI strategy by refusing to engage with AI tools. Meanwhile, a Kyndryl survey found that 45 per cent of U.S. CEOs believe most employees are resistant or even hostile to AI, he says.
At the same time, shadow AI is becoming more common, with employees secretly using unsanctioned tools without oversight or support. This creates security and compliance risks and prevents the development of shared best practices, says O’Connor.
Both trends stem from fear, mistrust, and a lack of agency, he says.
“For much of the workforce, their recent experience of supposed productivity-enhancing tools has been a bait-and-switch: more expectations and workload coupled with more clutter and distraction. They've been told technology would make their lives easier, and instead it made their phones buzz at 10 p.m. Without addressing this fundamental breach of trust and creating genuine shared gains, AI initiatives will continue to fail at the human level.”
The changing dynamic of efficiency
O’Connor believes that the future of work isn’t efficiency, but effectiveness: “It’s not that efficiency is irrelevant, but it’s no longer the human differentiator."
He explains that digital labour will always beat human labour on speed and volume.
“If agentic AI can infinitely scale an efficient, repeatable process, exerting immense efforts to squeeze an extra 1–2 percentage points of efficiency out of the human workforce becomes a pointless endeavour. In an AI-augmented world, human value creation has fundamentally shifted, as competing with AI on efficiency is a dead end.”
That means the workforce must focus on being effective, says O’Connor.
“The future of human performance lies in effectiveness—doing fewer, higher-impact things better, while AI handles the task volume.
“That means doubling down on the skills AI cannot replicate: creativity and innovation, strategic judgement with imperfect information, emotional intelligence and relationship building, critical thinking and analysis, and adaptability and learning agility. And these same skills are best leveraged by optimising the energy, recovery, motivation and wellbeing of the workforce.”
Most employers are still far from benefitting from the full potential of AI, according to a previous IBM report.
