Workflow Automation: Where to Start, and Where Most Teams Should

“Workflow automation” has become so broad a category that it’s almost useless as a planning concept. Some teams use it to mean Zapier rules. Some use it to mean custom-built RPA. Some use it to mean their CRM’s trigger system. The right starting point is the same regardless: find the workflow whose manual labor cost is highest, automate that one, prove the pattern, then expand.

Where most teams start (and why it’s wrong)

Most teams start by automating the workflow that’s easiest to automate. Lead routing. Slack notifications. Calendar reminders. These feel like wins and they are, modestly. But they don’t address the workflows actually costing the business meaningful money.

The result is a graveyard of small automations and a leadership team that says “we tried automation, it didn’t move the needle.” The automation didn’t move the needle because it wasn’t aimed at the needle.

Where to start (the honest answer)

Three questions to identify your starting point:

• Which workflow is run most often by your most expensive labor?
• Which workflow, when it goes wrong, costs you a customer or a deal?
• Which workflow generates the most internal coordination overhead (meetings, Slack threads, escalations)?

Whichever scores highest across all three is where you start. In our experience, it’s usually quoting, scheduling, or weekly reporting. Almost never lead routing.

The three layers of workflow automation

• Rules-based automation (Zapier, Make, Workato, CRM triggers) — cheap, fast, brittle. Right for high-volume, low-judgment steps. Wrong for anything with edge cases.
• Workflow tools with state (Asana automations, project management workflows) — better for multi-step processes with handoffs. Limited by the tool’s shape.
• Custom-built workflow systems — right when the workflow has business logic that’s yours, not generic, and the cost of manual operation is high enough to justify the build.

The mistake to avoid

The most expensive mistake in workflow automation is investing in the wrong layer. Trying to handle a complex business workflow in Zapier produces a brittle pile of rules nobody understands six months later. Building a custom system for what should be a simple Zap is a six-figure mistake.