
Create domain-specific prompts that work with your AI tool's actual reasoning mechanics
You'll have 5-8 tested prompts for recurring tasks in your work that you've refined based on what produces reliable outputs from your chosen model.
What you'll make
Domain Prompt Library
- Task Breakdown: reasoning requirements per task
- Working Prompts: tested versions with success rates
- Failure Patterns: where AI reasoning breaks and why
The 3 parts
Each part ends with a check you can see.
Map Your Task Reasoning Requirements
Break down 5-8 recurring tasks by what type of reasoning each requires from AI.
Done when
You have 5-8 tasks mapped by reasoning type with initial prompts written for each.
Test Where AI Reasoning Breaks
Run your initial prompts on real work and identify specific failure patterns.
Done when
You have tested each prompt 3+ times and documented specific failure patterns with examples.
Rebuild Prompts Around AI Reasoning
Redesign your prompts based on failure patterns to work with how your AI tool actually processes information.
Done when
You have rebuilt 5+ prompts that consistently produce usable output across multiple test runs.
Capstone: you put together your Domain Prompt Library.
About this lesson
What you'll be able to do
- Build 5-8 tested prompts for recurring work tasks that consistently produce usable outputs
- Judge when AI reasoning breaks down and needs restructuring rather than prompt tweaking
- Iterate prompts based on output patterns rather than trial-and-error
Who it's for
- Knowledge workers who use AI regularly but get inconsistent results
- Professionals who want reliable AI workflows for specific recurring tasks
- People frustrated by hit-or-miss AI performance who want to understand why
What it won't cover
- How AI models work internally
- Advanced prompt engineering techniques
- Building custom AI tools or fine-tuning models
Skills you'll practise
- Task decomposition for AI reasoning
- Output pattern analysis
- Iterative prompt refinement
- Reasoning failure diagnosis