Business Automation Starter Playbook
How to find the hours worth automating and which ones to leave alone.
How to do it
- 01
Log two weeks of admin
Every repeated task, who does it, how often, how long. Estimates are always wrong; logs are not.
- 02
Price each task
Frequency multiplied by duration multiplied by hourly cost. Now you can rank by money, not annoyance.
- 03
Score by judgement required
High frequency and low judgement is the automation sweet spot.
- 04
Automate one process end to end
One finished automation beats five half-built ones.
- 05
Run parallel, then switch
Keep the manual process for two weeks alongside the automation.
- 06
Add monitoring
Every workflow needs an alert on failure. Silent failure costs more than manual work.
- 07
Review quarterly
Processes drift. Automations should be revisited as the business changes.
The best first candidates
These come up in nearly every owner-led business and are safe to automate early.
- Instant response to new enquiries
- Appointment reminders and no-show recovery
- Quote and proposal follow-up
- Client onboarding paperwork and welcome sequences
- Invoice issue and payment chasing
- Review requests after completed work
- Weekly performance reporting to the owner
What not to automate
Complaint handling, pricing negotiation, anything requiring discretion, and any message that would embarrass you if it arrived at the wrong moment. Automation should carry the admin and hand the judgement to a person.
Designing for failure
Integrations break, APIs change, credentials expire. Assume it will fail and design so that failure is loud: alerting, logging, and a fallback that puts the task back on a human's list.
The honest return calculation
Hours returned multiplied by their real cost, minus the cost of the automation. If a workflow saves twenty minutes a month, leave it alone. If it saves six hours a week, it changes the business.