Prompts

Prompt: Generate Process Automation Specifications

Jay Banlasan

Jay Banlasan

The AI Systems Guy

tl;dr

Turn a business process into detailed automation specifications that a developer or tool can implement.

This prompt process automation specifications turns a messy business process into a clear spec document that a developer, AI, or automation tool can implement. The gap between "we should automate this" and actually automating it is almost always the specification.

The Prompt

You are a business automation architect. Turn this process into detailed automation specifications.

Process name: [e.g., new lead follow-up, invoice generation, weekly reporting]
Current process (describe it in your own words, as rough as needed):
[Describe what happens step by step, including who does what, what tools they use, and any decisions they make along the way]

Tools we currently use: [list all relevant tools and platforms]
Who currently does this manually: [role/person]
How often this process runs: [e.g., every time a lead comes in, weekly, monthly]
What triggers it: [what event starts the process]

Generate automation specifications:

1. PROCESS MAP:
Convert my description into a clean, numbered sequence of steps. Identify:
- Automated steps (no human judgment needed)
- Decision points (human judgment required)
- Data handoffs (information moving between systems)

2. TRIGGER DEFINITION:
What event starts the automation? Define:
- Trigger type (webhook, schedule, manual, data change)
- Trigger source (which system fires it)
- Data available at trigger time

3. FOR EACH AUTOMATED STEP:
- Input: what data comes in
- Process: what transformation or action happens
- Output: what data goes out
- Destination: where the output goes
- Error handling: what happens if this step fails
- Validation: how to confirm the step completed correctly

4. DATA FLOW DIAGRAM:
Text-based diagram showing data moving between systems:
[System A] -> (data) -> [Process Step] -> (result) -> [System B]

5. INTEGRATION REQUIREMENTS:
For each system involved:
- API availability (does it have one?)
- Authentication method
- Key endpoints or features needed
- Known limitations or gotchas

6. TESTING PLAN:
- Test cases for the happy path (everything works)
- Test cases for failure scenarios (what could go wrong)
- Sample data to use for testing

7. IMPLEMENTATION RECOMMENDATION:
- Suggested tool or approach (Zapier, Make, custom script, Claude Code)
- Estimated build time
- Estimated cost per run
- Maintenance requirements

Be specific enough that someone who has never seen this process could build the automation from this spec alone.

The Decision Points

Section 1's distinction between automated steps and decision points is the critical output. Most failed automation projects try to automate decision points. Keep humans in the loop where judgment matters. Automate everything else.

Building From the Spec

Hand this spec to Claude Code and say "build this." The specification is detailed enough to produce working code. That is the whole point.

Build These Systems

Ready to implement? These step-by-step tutorials show you exactly how:

Want this built for your business?

Get a free assessment of where AI operations can replace overhead in your company.

Get Your Free Assessment

Related posts