The strongest wins come from the ground level — not from an 18-month transformation program.
Every week, I meet leaders who want to "do automation" but don't know where to start. The good news: you don't need a massive project for measurable results. Organizations I support often find their first win in half a day of observation, not six months of planning.
At a glance
- Three signals: repetitive, consumes key people's time, produces visible errors
- Ask "If we freed five hours a week, what more valuable work would this person do?" before talking tools
- Prioritize by impact, feasibility, risk, and buy-in — not technical complexity
- A well-scoped pilot is often enough; "digital revolution" can wait
Three signals a process is a good automation candidate
1. It's repetitive and predictable
Same inputs, same steps, same outputs — week after week. Think data entry between systems, report formatting, follow-up reminders.
2. It eats time from key people
When a senior engineer, project manager, or partner spends hours on admin work, the real cost is high — well beyond the hourly rate. You're paying for judgment at mechanical-work prices.
3. Errors have visible impact
Missed deadlines, conflicting versions, forgotten follow-ups — these irritants cost client trust and mental load.
If a process checks all three boxes, it deserves a closer look — before talking about tools.
The question I always ask first
"If we freed five hours a week for this person, what more valuable work would they do for the business?"
That reframes the conversation when a vendor promises "10x productivity" with no target process. We're talking about recovered human value, not tech for tech's sake.
Five often-overlooked processes
- Moving data between email, Excel, and an ERP (enterprise resource planning) or CRM (customer relationship management) system
- Producing meeting minutes or status reports
- Manual reminders (deadlines, approvals, missing documents)
- Consolidating scattered information before a meeting or decision
- Updating status in multiple tools that don't talk to each other
None of these require a "digital revolution." A well-scoped pilot is often enough. At a Montreal law firm, standardizing phone conference notes — with human validation — freed the equivalent of half an admin role without touching legal judgment.
A half-day exercise that works
Gather three people who actually do the work — not only managers. For each candidate process, write:
- Steps as they happen today (not as the policy manual says)
- Time per week, even as a rough estimate
- What breaks when someone is absent
In most sessions, one process stands out within 90 minutes. That's your short list. Everything else can wait.
How to prioritize without drowning
| Criterion | Question |
|---|---|
| Impact | How many hours or errors avoided per month? |
| Feasibility | Is the process documented or at least observable? |
| Risk | What happens if we get it wrong the first week? |
| Buy-in | Does the team involved want it to change? |
Start with high impact and low risk. That's your first quick win — and your best case for what comes next. Avoid the trap of the "most technically impressive project" — it's not always the one that frees the most value.
What I see in SMBs that succeed
They don't try to automate everything. They pick one concrete irritant, measure before and after, and let the team validate the result. That's my approach: understand operations, identify friction, uncover quick wins — in that order.
Where you are
Too small for automation? dismissed the size myth; here you identify what to automate first. Next: When automation fails (it's rarely the technology), to avoid human and operational pitfalls.
If you're torn between several opportunities, Let's narrow it down — in 30 minutes we can often isolate one worth piloting.