“We had to make sure someone couldn’t just type into the chat, ‘What is my CEO’s salary?’” said Chris Blazejewicz, IT manager at Lakeside Process Controls. “That would be a resume-generating event for me.”
The line landed well with the crowd at ManageEngine UserConf Toronto. Funny, yes, but it hit a nerve. It captured the day’s underlying tension: without the right safeguards, AI doesn’t just help your team, but it can just as easily blow up your workflow.
For Blazejewicz and other IT leaders in the room, the question wasn’t whether to embrace AI. It was whether their infrastructure, governance, and culture were ready for what that actually means.
Throughout the event, which featured a keynote from ManageEngine CEO Rajesh Ganesan, a fireside chat with Blazejewicz, and a panel of IT leaders from across Canada, one message came through clearly.
AI is not a shortcut. It’s an accelerant. And it’s only as useful as the systems and people around it.
Getting to intelligence means finishing the earlier stages
In his opening keynote, Ganesan outlined four stages of IT evolution: systems of record, systems of workflow, systems of experience, and finally, systems of intelligence. Most organizations, he argued, are eager to jump to the last one.
“All of us want to be on the fourth stage today. We all want to be AI-ready,” he said. “But most people miss that they have not done the first three phases well. Maybe one phase well — it doesn’t still count.”
For those who have done the work, though, the payoff is starting to appear. ManageEngine showcased tools like Ask Zia, a conversational assistant, and Zia Agents, which Ganesan called “digital employees” that can act independently within IT environments.
Agentic AI models, like Zia Agents, where systems take initiative rather than simply respond, represent the kind of capability that becomes possible only after the foundational stages are in place. Far from being shortcuts, they’re what’s waiting on the other side of the hard work.
“Just throwing all the available AI is not going to solve your problems,” he said. “You need to have clarity. What do you want to achieve, who are your customers, and what is your business?”
What a real Copilot rollout actually looks like
That need for clarity played out in detail as Blazejewicz recounted his team’s Copilot deployment across Microsoft 365. While their leadership was eager, he urged caution, not because the tools didn’t work, but because they didn’t know what they didn’t know.
What followed was six to eight months of preparation. Governance layers were added. Data classification was overhauled. Expectations were reset.
One early challenge? Copilot’s classifiers repeatedly misread product stock keeping units (SKUs) as sensitive data.
“We had 50,000 South African social insurance numbers,” he said. “And we don’t even do business in that part of the world.”
On-premise data also proved difficult to integrate.
That led to a shift in strategy. The team opted to move selectively to the cloud, focus on tagging and ownership, and bring users into the process.
“Too often, everybody says data is an IT problem,” he said. “But half the time, we look at it and say, ‘We have no idea what this is.’”
The payoff came later. Once properly configured and paired with thoughtful prompts, Copilot began producing cleaner policy drafts and saving time on routine tasks.
Still, Blazejewicz was clear in saying that the point isn’t automation, but augmentation.
“These are tools to be used,” he said. “They’re not to replace us. They’re to help us.”
AI readiness looks different in every organization
That sentiment was echoed in the panel session, where IT leaders from Canada Cartage, Guelph General Hospital, and Abell Pest Control shared how they’re balancing modernization with the realities of their environments.
At Canada Cartage, director of IT infrastructure Alfio Costantino said his team is focused on getting more out of existing tools before bringing in new ones.
“We’ve got to pick our battles,” he said, emphasizing that the goal is to make systems work better, not necessarily make them new.
Colin James, senior director of digital health and IT at Guelph General Hospital, spoke about innovation under strict regulatory limits. While still using a 30-year-old health information system, the hospital was first in Canada to shift diagnostic imaging to the cloud, without compromising data residency.
“We will not place our data outside the country, period,” he said. “It’s a trust relationship.”
And Norm Waslynchuk, VP of technology at Abell Pest Control, offered a pragmatic view from a lean, in-house development team. His take on AI was simple: “It used to be called automation. It’s just a new hashtag.”
Despite their different sectors, all three agreed that true modernization isn’t about adopting flashy tools. It’s about aligning with purpose, understanding risk, and supporting the people doing the work.
The future belongs to teams that prepare now
While the technologies on display — from co-pilots to agentic AI — point to a more autonomous IT future, the message at UserConf Toronto wasn’t one of disruption, but discipline.
“You can have it done faster,” said Blazejewicz. “[Or] you can have it done right. What would you like?”
For organizations serious about leveraging AI, the next move isn’t experimentation for its own sake. It’s building the right foundations, so that when AI is ready to lead, your systems and people are ready to follow.
Digital Journal is a media partner of ManageEngine UserConf Toronto.
This article was created with the assistance of AI. Learn more about our AI ethics policy here.
