Canadian enterprises are racing to implement AI, but many are finding that the hardest work has nothing to do with algorithms. The real challenge lies beneath the surface, in the outdated systems and data architecture that fail to support it.
“AI is a very data hungry and data intensive process,” says Lukas Lhotsky, chief executive officer of Ateko, a Bell-backed IT services firm that connects and modernizes core enterprise systems. “Some are struggling with just a very foundational question around what their data ecosystem is going to be like, and how that data ecosystem is actually going to be stitched together so that they can take advantage of AI.”
Bell sponsors both the Canadian CIO of the Year Awards and the CIO Hall of Fame, which are run by the CIO Association of Canada. Lhotsky spoke with Digital Journal ahead of this year’s event to discuss the broader system challenges shaping the work of technology leaders across the country.
CIOs across Canada are navigating similar realities. They are being asked to modernize aging systems, streamline operations and introduce new digital capabilities, often while keeping essential services running without interruption.
Meeting these expectations requires coordination, clarity on priorities, and steady investment in core infrastructure to ensure underlying systems are ready for the change leaders want to drive.
The hidden costs of data fragmentation
This foundational challenge shows up across some of Canada’s most complex modernization efforts, including the work led by several of this year’s CanadianCIO award winners.
The Canadian Cancer Society has rebuilt its technology foundation to unify systems supporting donor engagement. Beneva confronted similar issues after merging two large insurers. Nova Scotia’s health transformation is progressing following the establishment of shared standards for data and workflows.
Each case reflects the same reality. Modernization requires coherent, accessible, and integrated information.
Many organizations lack a single, accurate view of their information. They have decades of data scattered across formats, business units and legacy applications.
Knowledge bases do not align. Systems of record do not integrate cleanly. Workflows often rely on custom logic built years ago that hasn’t been updated for current operating realities.
Lhotsky sees this as a foundational issue. AI applications, especially next-generation agentic systems, can only deliver value when they can seamlessly access a structured knowledge base.
As he puts it, AI only works when organizations can “go mine the underlying knowledge base, or the underlying code base that exists and have it be meaningful to some of the agentic work that is now expected at [the] enterprise.”
Why early AI experiments stall
A second major challenge is execution.
“We’ve all been seduced by the idea of agentic enterprise and of AI agents,” Lhotsky says. “They’re very good at certain things, but they’re not particularly good at getting things done.”
Many organizations have early agentic pilots. Few have connected them to the systems that process payments, validate identities, manage inventory or complete transactions. Without integration into these systems of action, AI remains a demonstration rather than an operational capability.
This gap widens when enterprises attempt end-to-end AI transformations.
Lhotsky points to tightly scoped workloads as the most reliable path to value. He describes an example from a Canadian retailer where an AI agent addressed loyalty point disputes by integrating with transaction systems.
Because the workflow was simple, the value was immediate. Large-scale, highly abstract initiatives, he says, “end up doing very little, because it’s just so vast and it’s so complex.”
Governance, guardrails and the return of simplicity
The third challenge is governance.
“In many cases, an old-fashioned deterministic if this, then that process is really good,” Lhotsky says, adding that even when a simpler workflow would be more reliable, “we’ve been seduced by this idea that an agent can do things.”
This tension is playing out across Canada. Western University recently strengthened its cybersecurity posture across a complex campus environment, a shift that earned Brent Fowles the CISO of the Year award and underscored the link between governance and modernization.
Lhotsky sees a pattern across enterprises. Some teams are using AI to govern AI, creating loops that are difficult to control. Others attempt experimentation without a defined business outcome.
Portfolio discipline becomes essential. Determining when to apply deterministic logic, when to deploy agentic systems and how to measure repeatable value is increasingly part of the CIO’s mandate.
The CIO balancing act
CIOs must modernize at speed while maintaining the stability of mission critical systems. As CEO of Ateko, Lhotsky draws on the company’s long focus on resilience.
“Our entire history at Ateko is managing mission critical infrastructure,” he says. “So the idea of operational resiliency is so fundamental to our identity.”
CIOs also face rising architectural complexity. API-driven systems create interdependencies that make failures harder to diagnose. Monitoring, traceability and observability have become core competencies.
At the same time, CIOs must push their organizations toward simpler, standardized platform models. Lhotsky describes this shift as “click, not code.”
“Developers like to develop things,” he says, adding that many organizations over-customize systems that already work for hundreds of global enterprises. “Where is something good enough? Where a baseline data model or a baseline workflow is good enough and can be maybe just dialed in, as opposed to being built entirely on its own?”
This shift changes the identity of IT teams. It also changes the role of the CIO, embedding them deeply in financial modelling, business outcomes and strategy.
“They’re reluctantly being forced into the business imperative conversations,” he says.
That shift will accelerate as agentic systems become more common.
The national imperative: Modernization and competitiveness
Lhotsky argues that CIO leadership is now central to Canada’s innovation economy, saying that “CIOs have a tremendous role to play.”
Part of that responsibility is talent. Canada produces strong technical graduates, yet many leave for roles abroad. CIOs can influence that trajectory by building IT environments that expose workers to modern, globally relevant platforms.
“They are able to create infrastructure that can be used and that can train entire generations of Canadians,” Lhotsky says.
He also believes government has an important role to play in shaping that environment.
“The ones that have been the most innovative have been the ones that have said, we’re going to use our purchasing power,” he says.
When governments act as early customers for new technologies, it can strengthen resilience, support talent development and create clearer pathways for domestic firms to participate in large-scale digital projects.
This view aligns with recent policy signals from Ottawa aimed at leveraging public spending to bolster domestic technology adoption and competitiveness. The recent federal budget introduced a renewed commitment to Buy Canadian, along with a proposal to expand procurement opportunities for small and medium-sized technology companies.
The Canadian Council of Innovators (CCI) welcomed the direction, stating that “government procurement is the most impactful way to scale Canadian companies” and that buying from homegrown innovators “creates anchor customers that help our scale-ups compete globally.”
CCI also noted that, paired with Prime Minister Carney’s Buy Canadian policy, the federal approach signals “a government that is meaningfully interested in collaborating.”
Private sector decisions matter as well.
“It’s incumbent on Canadians to buy from Canadians,” Lhotsky says, arguing that domestic companies can provide deep capability while keeping economic value in the country.
When organizations choose platforms and partners that reflect national priorities, they help build the technical foundations that the next generation of workers and firms will depend on.
For CIOs, he keeps coming back to the need to define the work before doing it.
“Scope,” he says. “Invest in context. Frame the problems clearly.” Clear outcomes, contained use cases and coherent systems allow AI to deliver value at scale.
The systems CIOs rebuild today will determine whether the country can turn AI optimism into long-term competitiveness.
Final shots
- AI projects stall not because of models but because underlying systems and data aren’t ready.
- Tight scoping and simple workflows often deliver more value than large end-to-end transformations.
- Governance matters. Deterministic logic is still the right tool for many enterprise workloads.
- CIOs need modern, standardized platforms to balance speed with operational resilience.
- Public and private buying decisions shape Canada’s digital economy and the talent it can retain.
Digital Journal is the national media partner for the CIO Association of Canada.
