Skip to main content
Skip to content
Menu

Agents Beginner 3 parts

Write a structured agent instruction file for your team's AI tool, then measure its impact on output

You'll have a 100–150 line instruction document (AGENTS.md style) plus a side-by-side comparison showing the quality difference when your AI tool uses it versus operating without guidance.

What you'll make

AI Tool Instruction File with Impact Analysis

  1. Tool Context: what AI tool this guides and what tasks it handles
  2. Core Instructions: the structured guidance for consistent output
  3. Domain Rules: specific knowledge your tool needs about your work context
  4. Quality Standards: what good output looks like in your domain
  5. Impact Evidence: before/after comparison with specific quality measures

Write your instruction file in a format your team can actually use, then document specific examples where it improved output quality.

The 3 parts

Each part ends with a check you can see.

  1. Map your AI tool's current behavior

    Document exactly how your team uses AI now and where output quality varies.

    Done when

    You have 5+ real examples of your team's AI interactions with quality scores for each, showing where consistency breaks down.

  2. Rank what expertise to capture first

    Decide which pieces of your domain knowledge will have the biggest impact on AI output quality.

    Done when

    You have a prioritized list of 8-12 domain knowledge gaps ranked by impact potential, with specific examples of how each gap currently hurts output quality.

  3. Run controlled tests of instruction effectiveness

    Test your instruction file against real tasks and document specific quality improvements.

    Done when

    You have side-by-side comparisons showing measurable quality differences when AI uses your instructions versus operating without them on the same tasks.

    Capstone: you put together your AI Tool Instruction File with Impact Analysis.

Where this helps

An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.

About this lesson

What you'll be able to do

  • Write a structured instruction file that measurably improves your team's AI tool output
  • Document specific quality differences with before/after evidence
  • Have a reusable method for testing instruction file changes
  • Know what guidance works vs what AI ignores in your context

Who it's for

  • Teams already using AI tools for work who want more consistent output
  • Anyone who's frustrated with AI giving generic responses to domain-specific tasks
  • People who need to onboard others to use AI tools effectively

What it won't cover

  • Building AI agents from scratch or technical implementation
  • Advanced prompt engineering theory or optimization techniques
  • Setting up new AI tool infrastructure or changing platforms

Skills you'll practise

  • Domain expertise capture in structured formats
  • Comparative output evaluation against quality standards
  • Controlled testing methodology for AI tools
  • Instruction refinement based on performance data