A 7-step playbook for founders and lean teams: how to audit your operation, score what's worth automating, and pick the first three workflows that earn their keep — without burning a quarter on the wrong project.
7
Decision steps
3
Workflows to pick
30 min
To complete
Unlock the playbook
Where should we send it?
Drop your email and the full 7-chapter playbook + the audit checklist unlock on this page. No drip campaigns, no agency follow-up sequence.
The playbook
Seven decisions before you automate anything.
Chapter 01
Map the operation in one page
Before you touch a tool, list every recurring workflow that consumes founder or operator hours. One row per workflow: trigger, steps, owner, frequency, time-per-run. If it isn't on this page, you can't measure it.
"If a workflow lives only in someone's head, automating it just hard-codes the chaos."
Chapter 02
Score every workflow on four axes
Volume (runs per month), Time cost (minutes per run), Decision quality (does humans-in-the-loop matter?), and Stability (does the workflow change weekly?). Multiply the first two for raw hours; use the last two as veto flags.
High volume × high time = automate first
Low stability = document before you automate
Decisions with reputational risk = assist, don't replace
Chapter 03
Separate research, decision, and execution
Most founders try to automate end-to-end. That fails. Break each workflow into three layers — research (gather context), decision (judge it), execution (do the thing). AI is excellent at layers 1 and 3. Founders own layer 2.
Chapter 04
Pick exactly three first automations
Not five. Not fifteen. Three. One research workflow, one execution workflow, one reporting workflow. This gives you proof, momentum, and a template — without burying the team in half-finished systems.
"Three working systems beat fifteen prototypes every time."
Chapter 05
Write the rollback before the rollout
Every automation needs an answer to 'what happens when this breaks at 11pm on a Friday?' Manual fallback, owner, and detection method documented before launch. No exceptions.
Chapter 06
Instrument from day one
Every automated workflow logs: runs attempted, runs succeeded, time saved per run, human overrides. Without instrumentation you can't tell a working system from a hallucinating one.
Runs attempted vs. succeeded
Minutes saved per run (validated, not estimated)
Override rate — when humans corrected the output
Last failure cause
Chapter 07
Review every 30 days, kill ruthlessly
Run a 30-minute review. Anything with override rate above 20%, or that hasn't saved meaningful time, gets killed or rebuilt. Lean teams can't afford automation graveyards.
The audit checklist
Print this. Tape it next to your desk.
1Workflow inventory built (one page, all recurring work)
2Each workflow scored on volume, time, decision risk, stability
3Research / decision / execution layers separated
4Top three automation candidates selected
5Rollback + owner documented for each
6Instrumentation defined (success, override, time-saved)
730-day review on the calendar
Want this run on your operation?
The Operator Sprint is the same audit, run directly on your business in two weeks — with the first automation built before we finish.