Thanks for subscribing to Digital Journal's What does this really mean? newsletter. We dig into the week's innovation and business news and try to explain what it actually means for the people running organizations. Follow us on LinkedIn and join the conversation. We're Canadian media, so Meta's not an option. Don't get Chris started. |

Veeam’s Jeff Reichard opens Day 2 of Peer Forum with Canadian AI governance data. — Photo by Jennifer Friesen, Digital Journal
A couple of things before we get into it.
You may have noticed this isn't arriving on a Friday afternoon. We're moving to Tuesday morning delivery. You may also have noticed that today is Thursday.
Last week, I spent two days at a conference full of CIOs. Inspired, I came home and promptly overhauled our marketing tech stack. Turns out talking to CIOs is not the same as being one.
But the dust has settled. We now have a newsletter archive: newsletters.digitaljournal.com. You can now catch up on missed issues or share them with colleagues. You could also use them as party trivia, but if AI policy is your go-to social lubricant, we both need better invitations.
Now, about that conference.
Early on the first morning of the CIO Association of Canada's annual Peer Forum in Vancouver, Lenovo's John Encizo asked the room a simple question. How many of them have given AI agents employee IDs?
One hand went up.
I'll admit I hadn't thought about whether AI agents might need employee IDs. Maybe those who didn't put their hands up had, but just hadn't done it yet.
Either way, it felt like a good example of why they gather at Peer Forum.
Every CIO in that room is navigating the same fog of constant change at increasing speed.
The City of Kelowna is running AI chatbots on the same infrastructure as code from 1999. CIBC compressed some workflows from 10 hours to 10 minutes after putting its entire through AI training before deploying anything.
The annual gathering is a chance to compare notes on what’s working, what’s failing, and how to find a path through together.

The fog has a name
Lenovo's John Encizo brought Canadian data showing 67% of employees are already using AI tools their IT teams didn't sanction.
Info-Tech's Geoff Nielson, drawing on research across more than 25,000 IT leaders, noted that three-quarters of CEOs think their IT function is in firefighting mode, when they'd rather it was thinking strategically.
Veeam's Jeff Reichard cited Gartner's projection that 60% of AI projects will fail this year because the organizations around the technology weren't ready for it.
These numbers came up across both days, in different sessions, from different speakers. I don't think any of them were surprising to the people in the room.
That's the thing about fog: you know you're in it.
The people finding their way through
The way out of the fog is to treat these obstacles as data points.
Take the shadow AI problem. If two-thirds of employees are routing around IT to get work done, Encizo argued that's a clear signal about your internal tools and processes.
"Don't be the ministry of no," he said. "They're going to use it anyway. Get ahead of it and give them sandboxes."
Arctic Wolf worked 1,000 incident response cases last year, and CISO Adam Marrè walked the room through the findings. In 65% of them, attackers got in through remote access tools the organizations had set up themselves.
The AI-driven attack wave is coming, but right now attackers are just logging in. Default credentials, dormant accounts, old breached passwords. Maintenance and discipline are the defence.
Zoho's Raju Vegesna pushed past the operational questions entirely. His argument used coffee as an example. Colombia grows the best beans, Italy built the culture and captured the value.
Canada produces an energy surplus, and energy is the key input for AI infrastructure. We have the energy, the talent, and the technical capacity. The question is whether we build the industry or keep supplying the raw materials for someone else to.
"History remembers builders, not suppliers," he said.
The Watercooler
Some light reading for when you're trying to plan something and the world keeps updating its terms.
Renting intelligence is a risky way to run a business — Dr. Andrew Forde of KPMG Canada takes the AI-as-utility argument seriously enough to dismantle it. If intelligence becomes metered infrastructure controlled by a handful of providers, decision-making becomes capacity-constrained.
Canadian tech companies are already global. Now they want to be indispensable — A Council of Canadian Innovators survey of 125 Canadian tech CEOs finds their top priority is building a customer base, not accessing capital or navigating regulation. They want their technology embedded as the standard others build around. Whether government policy is keeping up is a separate question.
Val Kilmer returns via AI as filmmakers test Hollywood's red line — A digital version of Kilmer, recreated with his family's consent and access to his personal video archive, appears in a trailer for a film he'd been cast in before the pandemic. The filmmakers followed SAG-AFTRA's three Cs: consent, compensation, and collaboration. Whether that framework holds as the technology gets cheaper will be worth watching.
Click and send this story: Email is 55 years old — Ray Tomlinson sent the first email in 1971, chose the @ symbol to separate the user from the machine, and when asked about it later said it "seemed like a neat idea." Fifty-five years later, and some people haven’t seen inbox zero since.
Final shots
You don't go to a conference like the Peer Forum for the answers. You go to find out who else is asking the same questions.
The same statistics that describe the fog are also the path through it. The conversations happening around them, about what's worked, what's failed, and what's worth trying next, are how leaders start to move.
One hand went up. It's not a lot, but it's a starting point.
Talk to you Tuesday.
— David
