Prompt Engineering for Operators
Not prompt theory. The structural patterns that produce usable output for the jobs you actually do — and a personal prompt library you'll use every day. 2 weeks. Get started — join Bodega Pass
Who this course is for
Operators who use AI regularly but whose outputs often miss the mark — or who waste time iterating before getting something usable. This course is the structural upgrade: after it, you will know exactly why a prompt produces a useful output or a mediocre one, and how to fix it fast.
What you'll build
A personal prompt library — 20–30 tested, reusable prompts for the specific jobs you do most often. By the end of Week 2, you can run them immediately without starting from scratch every time.
Course structure
| Week | What you'll learn | What you'll build |
|---|---|---|
| Week 1 | The 5 structural patterns that consistently produce usable output: role assignment, context loading, output format specification, constraint injection, and chain-of-thought triggering. Why each works, and when to use each. | 5 prompts — one for each pattern — applied to a real work task you choose |
| Week 2 | Application by job category: research and synthesis, writing and drafting, analysis and decision support, summarization, and conversational prompts. The specific patterns that work for each job type. | A personal prompt library of 20–30 tested prompts organized by your work categories |
Job categories covered
- Research and synthesis — company research, topic summaries, competitive analysis, and information extraction from documents
- Writing and drafting — cold emails, proposals, content, internal communications, and job descriptions
- Analysis and decision support — evaluating options, summarizing data, and structuring decisions
- Summarization — meeting notes, long documents, support tickets, and call transcripts
- Conversational and interactive — building assistants you use in repeated conversations with maintained context
Model coverage
Every prompt pattern is tested in ChatGPT (GPT-4 and GPT-4o), Claude (Sonnet and Haiku), and Gemini (Pro). The course includes notes on which patterns produce better output in which model — because the differences are real and matter for high-frequency tasks.
What's included
- Video lessons — one per structural pattern and one per job category
- 50-prompt starter library — pre-built prompts across all five job categories as a foundation for your personal library
- Prompt audit framework — how to diagnose why a prompt produces weak output and what to change
- Model comparison notes — patterns where ChatGPT, Claude, and Gemini produce meaningfully different results
- Community access — share prompts, compare outputs, and get feedback on patterns that aren't working
Time commitment
2 weeks. 2 hours per week. Self-paced. Short enough to fit into any schedule; valuable enough to return to when you take on a new type of work.
Common questions
- What AI tools do I need?
- A ChatGPT, Claude, or Gemini account — whichever you already use. A paid plan is recommended for access to the latest model versions, but the course works with free tiers.
- Is this about prompts for agents and automation?
- Partially — the patterns you learn are the foundation for agent and automation prompt design. The primary focus is prompts you run manually in a chat interface. For building agents, take Build Your Own Agent after this course.
- How is this different from the prompt packs sold in the store?
- Prompt packs give you tested prompts for specific use cases — fast value for a known job. This course teaches you how to build prompts yourself, diagnose weak output, and adapt patterns for any job. After this course, you can build your own packs for whatever you need.
Join Bodega Pass — includes every Academy course plus the full product library and prompt packs.