CRM Workflow Automation: 9 Practical Examples for Small Teams
Workflow automation is most valuable for the small, repetitive steps your team forgets on busy days. Here are nine automations you can set up in an afternoon, with triggers, conditions, and actions spelled out.
AI Custom CRM Team · · 2 min read
Automation in a CRM is simple in principle: when something happens (a trigger), check some conditions, then do something (an action). The value comes from choosing the right moments. The best candidates are steps that are easy to forget, happen many times a week, and follow the same rule every time.
Here are nine examples that work well for small teams. Each one is written as trigger → condition → action so you can adapt it.
Lead handling
- Instant assignment — New lead created → source is the website form → assign to the next rep in rotation and create a “call within 1 hour” task.
- Stale lead nudge — Lead status is still “new” after 24 hours → lead has an owner → notify the owner and their manager.
- Disqualification cleanup — Lead marked “unqualified” → reason is “not a fit” → remove from follow-up sequences and tag for quarterly review.
Sales pipeline
- Proposal follow-up — Deal moves to “Proposal sent” → no activity for 5 days → create a follow-up task and draft a check-in email for review.
- Won deal handoff — Deal marked “Won” → always → create onboarding tasks for the delivery team and notify them in your team chat.
- Large deal alert — Deal value updated → value above your threshold → notify the owner of the business.
Service and scheduling
- Appointment reminder — Appointment is 24 hours away → status is “scheduled” → send a reminder email or SMS with a reschedule link.
- No-show recovery — Appointment marked “no-show” → first no-show for this client → send a friendly rebooking message; on a second no-show, flag for staff review instead.
- Review request — Job marked complete → customer has an email address and no complaint is open → send a review request two days later.
Where AI fits in
Rule-based automations like these don’t need AI at all, and they’re easier to trust because you can read exactly what they do. AI becomes useful when a step involves judgment or writing: summarizing a long email thread before a call, scoring a lead against criteria you describe in plain language, or drafting a follow-up that a person reviews before sending.
A sensible pattern is to let rules decide when something happens and let AI help with what to say — with a human approving anything that goes to a customer.
Five rules for automations that stay useful
- Name every automation by what it does, not who built it.
- Start with notifications and tasks before automating customer-facing messages.
- Add conditions that prevent loops, such as “only once per record.”
- Check the run history weekly for the first month to catch surprises.
- Review the full list every quarter and switch off anything nobody would miss.
Getting started
Pick the one step your team most often forgets and automate that first. Once it runs reliably for a couple of weeks, add the next. A handful of well-chosen automations will save more time than dozens of clever ones that nobody remembers setting up.