1. List everything that runs
Keep one table of every automation: name, trigger, what it touches, who it affects and when it last ran. Most people are surprised by how many they have forgotten.
2. Read a sample of real runs
Open five recent runs of each important workflow and read the actual inputs and outputs. You are looking for drift: replies getting longer, categories getting vaguer, facts that are no longer true.
3. Check the facts it relies on
Prices, links, policies, opening hours and product names live in prompts and reference docs. Update them in one place and confirm every automation reads from it.
4. Check cost against value
Look at what each automation costs per month in tool and model usage, and how much time it saves. Turn off anything that costs more attention than it saves.
5. Check access
Remove connections the automation no longer needs. An old workflow with access to your inbox and payment tools is a risk even when it is switched off.
Before you call it done
- Inventory of every automation and its last run
- Five real runs read per important workflow
- Reference facts updated in one place
- Cost compared with time saved
- Unused connections and keys removed
Estimate the hours and money an automation gives back each month, and how fast it pays for the time it takes to build.
A free master guide and resource pack for people starting with AI automation: the systems behind voice agents, lead workflows and a small automation practice.
Get it free on Gumroad