Failure-First Automation Design

What it is for

Design what happens when a step fails before designing what happens when it works. An automation without that runs silently wrong for a week, which is worse than one that stops.

The skill Copy

Automations are easy to build and quiet when they break. Design the failure
path first.

Before adding any action, answer:
- **What happens if this step runs twice?** Triggers do re-fire. Any step that
  creates, charges, sends or posts needs an idempotency key or a lookup-first
  filter — otherwise a retry sends the customer a second invoice.
- **What happens if this step fails halfway?** If steps 1-3 wrote data and
  step 4 failed, the record is now half-made. Either make the last step the
  only one that commits, or add a cleanup path.
- **Who finds out?** Zapier's own error email goes to the account owner and is
  routinely filtered. Add an explicit error path that posts somewhere a human
  reads.

Then build:
- Put a Filter immediately after the trigger, not later. Every step after an
  unfiltered trigger runs on data you did not want.
- Use Formatter to normalise dates and numbers ONCE, right after the trigger,
  rather than in each downstream step.
- Name every step for what it does to the business, not what app it calls:
  "Create invoice", not "Zoho Books".

How to install it Copy

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from