Hi {{first_name}},

Most companies can tell you exactly how many office chairs they own. Ask them to list the software running the business, every program, who built it, and what it still does, and you might not get the same confident answer. 

In many cases, that list has grown organically, and with the rise of SaaS applications, IT hasn’t been the same gatekeeper it once was. 

Faced with this, Alberta fed its software through AI to read the code and build an inventory. Jennifer Friesen reported this week on what that information means for the people who buy technology. 

Once it had a clearer view of what it was running, the province says it could assess vendor pricing more accurately. They point to one project that was quoted at $54 million over three years, which they say is now being built in-house for about $2.5 million.

Knowing what you're running lets you challenge a quote. Friesen also points out what the AI can't do or take into account: the integration, the training, the approvals, and the person who's kept a system alive for 15 years without writing any of it down. 

Alberta says that the clearer picture is already changing how they interact with vendors. 

"We are a more sophisticated customer today than we were two years ago," said minister of technology and innovation Nate Glubish.

→ Read Jennifer Friesen's full story on what Alberta's numbers mean for IT pricing

My colleague Jennifer Kervin wrote about a governance panel at the Council of Canadian Innovators summit. Two things caught my attention. 

The first is the message you’re sending, intentionally or not, to employees when work is handed to AI. 

"The implications there are: I'm expendable. My job is training the bot that's going to replace me," said Kristina McDougall, who's spent more than 25 years placing leaders. 

The second is that the board is usually the last to know. Word about what a rollout is doing to staff, or their data, usually reaches directors through the CIO or CTO, if it reaches them at all.

→ Read Jennifer Kervin's full story on what your board isn’t hearing

Some light reading for when someone asks you to put a number on it

Canadian leaders trust their AI more than they can track it. A Deloitte Canada survey of 300 senior leaders found 95% of them confident in their AI compliance, and 64% conceding that at least 11% of their AI activity runs outside anything they've approved. What's that saying about managing what you can't measure?

Canada is investing in AI but trailing the US on returns. RSM surveyed 203 Canadian firms and found 43% said the investment beat expectations, against 57% of their U.S. peers. When Canadian pilots struggled to scale, data quality, cited by 53%, was the barrier named most often.

StatCan's first AI workforce survey finds AI still has to prove its place, reported here by Friesen, found 35.9% of workers had used generative AI on the job in the past year. A quarter of those users applied it to almost none of their tasks. And 56% of the people not using it said the tools don't apply to their work, which is harder to fix with more training.

Cyber incidents expose leadership gaps, not just IT issues. The Calgary law firm BD&P argues a breach tests a company's leadership as hard as its technology. The firms that come through best, they write, are the ones who settled who leads the response and who gets called before anything broke. Most boards can name their firewall vendor. Fewer can say who holds decision-making authority once an incident is underway.

Canada's real-time rail could transform payments, competition, and economic growth. In a break from explicitly AI news, Tim Sandle walks through how this new plumbing, which lets money move between banks instantly and around the clock, pairs with open banking to give fintechs a fair shot at the incumbents. Payments Canada figures it could produce about $3 billion in efficiency gains in its first five years.

For a long time, running technology came with automatic authority. You understood the servers, so everyone else had to come to you.

Then, the tools got easy at the same moment they got powerful. Anyone in the building can now ask an AI to write the code, redraw the process, draft the policy. What they can't always do is tell whether the answer is any good, and that part still rolls uphill to the person who used to just be seen as keeping the lights on and resetting passwords.

The job turned from running the machine into leading a whole organization's relationship with it, right when the machine started changing every few weeks. 

Same title. 99 new problems, and the servers ain't one.

— David Potter

If you have any questions, you can reply to this email. I read everything.

Do you have a story we should chase, or a tip you'd share with a peer?
Contact us here.

Know someone who should be reading this? Forward it on, or
they can sign up here.

This issue is presented by Calgary Economic Development.
If you want your brand in front of the people running technology across Canada, you can sponsor a future issue.