Which process to automate first, and which ones to skip

Start with the work that eats your week

The right first automation is not the one that looks easiest. It's the one that costs you the most time and follows a clear rule every time.

Think about your last month. What task did someone on your team do over and over, the same way each time, and complain about on Friday? That's your target.

A dental clinic in Kelowna spends 90 minutes every Monday morning entering patient cancellations into three different systems. A property manager in Surrey copies tenant enquiry details from email into a spreadsheet, then into the lease database, then into the accounting software. A 12-person accounting firm in Burnaby manually files 40 to 60 invoices each week into folders by client and date.

Each of these tasks: - Happens the same way every time - Takes 5 to 15 hours a week - Doesn't require a decision - Frustrates the person doing it

These are your first candidates.

The math that matters

Before you pick, count the actual hours. Not guesses. Track one person for a week, or ask them to time three cycles of the work.

If the task takes 2 hours a week, it's not worth automating. If it takes 8 hours a week and a junior staff member does it, that's worth $200 to $300 a month in reclaimed time. If a senior person does it, that's worth more.

But the real win is not the salary cost. It's the time your team gets back to do work that matters: client calls, problem-solving, work only they can do.

What not to automate yet

Be honest about what you're not ready for.

Skip anything that requires judgment. A plumber in Richmond gets 20 service calls a week. Some need a follow-up visit, some don't. That decision lives in the plumber's head. Automating it wrong costs you callbacks and angry customers. Leave it alone.

Skip work that happens once a month or less. If someone does a task 2 or 3 times a month, the setup time to automate it is longer than the payoff. Wait until you have 10 or 20 cycles to learn the real pattern.

Skip anything that touches a client relationship directly. A distributor in Calgary sends custom quotes. Each one is slightly different. A client expects to talk to a real person about price and terms. Automating the first email to a new prospect might work. Automating the quote itself will feel cheap and cost you deals.

Skip work where the rule changes. If your process changes every quarter, or if every client needs a different approach, automation will break and frustrate you. Wait until the rule is stable.

Why the first one matters most

Your first automation will teach your team how this works. It will show them what's possible without the drama. It will build trust.

If you pick a task that's messy, full of exceptions, or touches client relationships, it will fail quietly or create more work than it saves. Your team will say "AI doesn't work for us" and you'll have a harder time getting them to try the second one.

If you pick something simple, clear, and repetitive—something that genuinely wastes time—it will work. Your team will feel the relief on day one. They'll ask what's next.

Where to start

• Write down three tasks your team does every week the same way, and count the actual hours. Ask the person who does it, not the person who thinks it happens. Pick the one that takes the most time and frustrates someone the most.

• Check the rule. Does it follow the same steps every time. Are there exceptions. If there are more than two or three exceptions, it's not ready yet. If the rule is stable, move forward.

• Talk to the person who does the work. Not your manager instinct. Not what you think should be automated. Ask them what they hate most, what takes the longest, and what they'd rather be doing instead. That answer is usually the right one.