“Which job do I hand off first?”

“Entrepreneurs know they need AI, but many don’t know where to start.”

That’s Jean-Sébastien Charest, Chief Information Officer at BDC (the Business Development Bank of Canada), and I think he has it exactly right.

On 2026-11-02, BDC launched Mission AI under a headline I wish I had written: “Less AI hype, more results.” The plan is to find out what small and midsize businesses (SMEs) need, match those needs with Canadian AI tools, and have a panel of entrepreneurs test those tools in real businesses.

BDC’s numbers show why this matters. 69% of Canadian SMEs use AI very rarely or not at all. Only 5% have it deeply integrated into their operations.

Here’s the thing: I don’t think that 69% has a tool shortage. There are more AI tools than anyone could ever test. It’s like walking into a hardware store with no project in mind. Every tool looks useful, and you leave with nothing (or a table saw you never plug in).

The better question is much smaller: which job do I hand off first?

My current approach looks like this:

1. Pick one job your team repeats every week. Quotes, invoices, and sales follow-ups are good candidates.
2. Measure it. A job that takes one hour a week is worth about 50 hours a year. Now it has a dollar value.
3. Test one tool on that one job. A person reviews every output, because humans must remain in the loop (HITL).
4. Keep the tool if the number improves. Drop it if it doesn’t.

To be fair, BDC also reports that SMEs using generative AI are 24% more productive than those that don’t. That’s worth knowing, but it isn’t a promise. BDC says its survey results aren’t statistically projectable, and a gap like that doesn’t show AI caused the gain. Your own one-job test will tell you more.

I’ll be watching what BDC’s panel finds. Real test results beat another polished demo.

So what are your thoughts? Which job would you hand off first?