What Business Processes Should You Automate First?

"We should automate more" is a feeling. "Automate this process first" is a decision — and picking wrong costs real money, because the first automation project sets your organization's belief about whether automation works at all. Here is the selection framework we use with clients, and the candidates that win it almost every time.

The three-question filter

• How often does it happen? Automation is a fixed cost divided by repetitions. A daily task with ten touches beats a monthly report, every time.
• What does a miss cost? A missed lead is lost revenue; a missed reminder is a lost repeat customer; a mistyped invoice is a dispute. High miss-cost tasks justify automation even at low volume.
• Is the rule already clear? If your team can write the rule on an index card — "every inquiry gets a reply within a minute," "oil changes come due at ninety days" — it automates cleanly. If every case needs judgment, automate the routing to the human, not the decision.

Score your candidates on those three. The winner is almost never the flashiest process — it is the boring one that happens constantly, fails expensively, and follows a rule everyone already knows.

Candidate #1: answering your own leads

The most common expensive gap we find. We once audited a contractor's website whose "free estimate" form showed customers a friendly success message — and sent the inquiry nowhere. It had never been connected to anything. Nobody knew, because the failure was invisible by design. Even wired-up forms usually just drop mail into an inbox where response time depends on who checked it last. Instant acknowledgment to the customer plus routed alerts to your team scores maximum on all three questions — which is why we productized it as the Lead Rescue Sprint, live in five business days.

Candidate #2: the follow-up nobody remembers

Reminders, renewals, re-bookings — revenue that depends on someone remembering. The auto shop running our CRM platform automated exactly this: service reminders scheduled from real repair history, sent on a daily cadence, and the owner reports retention up 89% because of it. Frequency: constant. Miss cost: a customer who drifts to the shop that texted them. Rule clarity: an index card. Perfect score.

Candidate #3: re-keying between systems

Anywhere a person copies data from one screen into another — orders into accounting, bookings into a spreadsheet, leads into a CRM — is automation waiting to happen, with the bonus that machines do not transpose digits at 4:45 on a Friday. The integration does need engineering discipline (reconciliation, not blind copying — see our CRM integration guide), but the payoff is permanent.

What NOT to automate first

• A process that is currently broken. Automating chaos gives you faster chaos. Fix the process on paper first — that is the entire purpose of our $250 workflow map.
• Judgment calls. Pricing exceptions, sensitive customer conversations, hiring. Automate the paperwork around the decision; leave the decision human.
• The annual thing. If it happens twice a year, a checklist beats software.

Start smaller than feels ambitious

The best first automation is embarrassingly small, ships in days, and works every single time. Trust built, the second one gets easier to choose — and by the third, your team is bringing you candidates. That compounding is the actual product. When you want a partner for it, that is our workflow automation practice.