Somewhere in Canada right now, a meeting is wrapping up with the words “we should be doing something with AI,” and nothing is going to happen. Not because anyone disagrees. Because nobody in the room knows what the first step actually looks like.
Here is what it looks like. Five steps, no jargon, written for the owner of a business with 5 to 100 people, not for developers.
Step 1: Find where the hours actually go
Forget “AI strategy.” Walk the floor, or your inbox, and list the tasks that eat hours every week and make nobody happy. In most small businesses it is the same short list: writing quotes, answering the same inquiries, chasing invoices, retyping the same information into two or three systems, and assembling reports by hand.
Write the list down with a rough number of hours per week beside each. That number times a year of wages is what the task costs you. This is the entire foundation of business process automation with AI: know what the boring work costs before you spend a dollar automating it.
Step 2: Pick one task, not a transformation
The businesses that get burned on AI are the ones that start big. The ones that win start with one boring, expensive task and automate it well. Good first candidates share four traits: they are repetitive, they follow a mostly predictable pattern, they happen often, and getting them slightly wrong is annoying rather than catastrophic.
Quotes are a classic. So is inquiry triage: reading, sorting and drafting replies to the email that piles up overnight. So is data entry between systems that do not talk to each other.
Step 3: Keep a human in the loop
The goal is not to remove people. It is to move them from doing the task to approving the result. AI drafts the quote, a person clicks send. AI sorts the inbox and drafts replies, a person handles the three that actually need judgement. This is how you get the speed without betting the business on a machine being right every time.
Step 4: Measure it in dollars
Before the automation goes live, note the baseline: hours per week, response time, error rate, whatever the task’s honest numbers are. Thirty days after it ships, compare. If the automation is not clearly paying for itself, fix it or kill it. AI earns its place in a small business the same way an employee does.
Step 5: Then do the next one
One working automation changes the conversation. The team stops asking whether AI is hype and starts pointing at the next slow thing. That queue, one boring task at a time, is what an AI roadmap actually looks like for a small business.
What this costs in Canada
Honest ranges: a proper look at where AI would pay off in your business, an AI readiness assessment, should be free or cheap, and should end in a written list with dollar figures. A single automation build is typically a fixed quote in the mid four figures to low five figures. Ongoing tuning runs as a monthly retainer. Be suspicious of anyone who cannot give you a fixed price for one defined automation.
The first step is smaller than you think
You do not need an AI strategy. You need step 1: the list. If you would rather have a second set of eyes on it, that is exactly what our AI automation consulting starts with: a free AI audit that tells you, in dollars, what is worth automating and what to skip. Worst case, we tell you AI is not worth it for you yet. That is a real answer too.
Want the dollar figure for your own week? Run the boring-task math. We pre-loaded the data-entry example from this article; swap in your own numbers and send Simon the list.
