Multi-Step AI Workflow

You are an AI workflow architect. Design a multi-step AI workflow to automate [TASK].

Task to automate: [DESCRIBE FULLY — what goes in, what should come out]
Current manual process: [HOW IT'S DONE TODAY]
Tools available: [AI MODEL + TOOLS: e.g. GPT-4 API + Python / Claude + Zapier / Make.com]
Data inputs: [WHAT DATA FEEDS THE WORKFLOW]
Output format needed: [WHAT THE FINAL OUTPUT LOOKS LIKE]
Volume: [HOW MANY ITEMS / HOW OFTEN]

Workflow design:
1. Input ingestion: how data enters the system
2. Pre-processing: cleaning or formatting before the AI step
3. AI prompt step(s): the specific prompt(s) at each stage (with variables shown)
4. Output parsing: how to extract the structured result from AI output
5. Quality gate: how to catch bad outputs before they cause problems
6. Output delivery: where results go (email / spreadsheet / database / API)
7. Error handling: what happens when a step fails

Include a flowchart description and estimated time savings vs manual process.