Why AI across the employee lifecycle demands strategy, not just tools

Artificial intelligence (AI) is reshaping how Canadian employers recruit, onboard, develop, and retain workers. But as adoption accelerates, a troubling pattern is emerging: organizations are deploying AI in isolated pockets rather than as part of a coherent, lifecycle-wide strategy – and employees are bearing the cost, according to Bruce Weippert, Certified Human Resources Executive (CHRE) and President of TAP Strategy & HR Consulting in Toronto. 

Weippert says he works primarily with small and medium-sized businesses and sees the same pattern across many of them in their approach to implementing AI in their human resources (HR) practices. 

"AI has been rolled out or is being rolled out in silos, but that's not the experience that employees have," says Weippert. "You might have a specialized AI tool for your recruitment, another one for your onboarding processes, another one for training and learning and development – but then you land on performance management, which tends to be a very manual process, where all that great information that you have in all these other systems isn’t available to a manager." 

The consequence is that managers conduct performance reviews based on subjective impressions, disconnected from data gathered at every earlier stage of the employee journey, according to Weippert. "The data is fragmented, and I think that that's creating major challenges for integrating a holistic HR lifecycle," he says. 

Uneven AI adoption 

This finding is reinforced by ADP’s Canada Workplace Trends for 2026 report, which surveyed 1,008 Canadian businesses in late 2025. While 47 per cent of businesses believe AI can enhance skills development, only 19 per cent are currently using it for that purpose. Forty per cent agree that AI can assist in onboarding and offboarding, yet adoption across those stages remains significantly lower than in recruitment. 

Weippert says that many organizations lack the internal expertise to connect disparate systems. "HR people are not necessarily technology people," he says. "And so for them to be able to launch it on their own without that external expertise and a bigger vision, and building transparency and governance around it, it will always remain a fragmented system." 

Strategy before implementation 

Carina Vassilieva, Chief Human Resources Officer at Apotex in Toronto, places the solution squarely in strategic planning. Before any AI tool is deployed, the organization must have a clear view of what it is trying to achieve – and that view must be shared across functions, she says. 

"Have a holistic, strategic overview of what the strategy is and how different components of this strategy interact together," Vassilieva, who calls herself a “big believer” in data. "If you look at different pieces in isolation, this is where the risk is." 

She stresses that it’s important to understand from the beginning what the organization is trying to achieve. "Outcomes need to be clearly outlined and key performance indicators (KPIs) need to be put in place to see if we're tracking towards those outcomes," she says. "Strategy, outcomes, KPIs, metrics – it all needs to be outlined in advance in order to progress towards the intended North Star, not just do things – data tells the story." 

Vassilieva also emphasizes cross-functional engagement from the outset. "It needs to be enterprise strategy with different parts," she says. "And even if it's piloted in one function, it's important to involve cross-functional stakeholders to weigh in." 

The governance and compliance gap 

New compliance obligations are making governance more urgent. For example, as of January 2026, Ontario employers with 25 or more employees are required to state in publicly advertised job postings whether they use AI to screen, assess, or select applicants. The ADP report found that 70 per cent of Canadian businesses still don’t have an AI ethics policy, while nearly half – 46 per cent – say ethical management of AI is a priority but only 22 per cent have established a formal policy to govern it. 

Weippert says he sees organizations regularly present AI-generated employment agreements that are legally insufficient, missing key elements of legislation and case law around enforceability. "We see clients come to us all the time with employment agreements and policies that were created in AI, and they're not ones I would recommend that they give to their employees," he says. 

The responsibility rests firmly with the human reviewing the output, according to Weippert. "That's where a human set of eyes with knowledge and experience come in, because you need to know how to ask the right questions of AI,” he says. “Otherwise, you're going to read what it kicks out and you're going to think that's okay because it looks really good." 

Keeping humans in the loop 

Both Weippert and Vassilieva believe in the same principle around AI use in HR: AI should augment human judgment, not replace it. 

"It's very important for organizations to have an HR strategy that isn’t immediately focused on reducing HR headcount," says Weippert. "Employees still expect to be treated as individuals, and AI shouldn’t replace that individual touch." 

The ADP data reinforces this position. Eighty per cent of Canadian organizations agree on the importance of maintaining human oversight in AI applications, yet 38 per cent report having employees who are fearful of being displaced by AI. That fear, left unaddressed, carries a direct cost to engagement and trust. 

“AI can be used to provide information, gather data, help with decision-making without replacing the human touch, but we don't want to remove the human judgment, decision-making, planning, and all the great things that people bring to the table,” says Weippert. “It's not just good enough to put a bunch of AI processes in places saying, ‘We’ve washed our hands of it and we’ve got it done.’” 

Maintaining a vision as AI increases its presence 

Vassilieva frames her philosophy in terms of proactive risk management embedded from the very first stage. "Risk management should be proactive and strategic," she says. "And definitely the first stage is most important and implementation follows, as long as you know what it is that you want." 

As agentic AI – systems capable of autonomously completing multi-step tasks – begins to surface across HR functions, the stakes of getting this foundation right will only grow. Chief human resources officers (CHROs) anticipate a 327-per-cent increase in AI agent adoption within their organizations by 2027, according to a 2025 Salesforce survey of global HR executives. 

“It's really critical that HR ensures there’s a sharing of information in a meaningful way throughout the organization of the data that they hold,” says Weippert. “The most important stage is to make sure that you’re building a clear vision for AI adoption in HR and what it is that you want it to truly do for you.”