Topic 2.3

Clear Explanations Are Hard to Write. That's the Point.

AI drafts at any complexity level. You make it accurate.

⏱️ 12 minutes 📋 Prompt Templates ✓ Quality Checklist

The Mental Load

Explaining complex concepts clearly takes effort. Choose the right level of detail. Structure logically. Add examples that clarify, not confuse. Use language your audience actually understands.

Time to write one solid explanation: 45-60 minutes.
Time to write five for a module: most of your day.

💡 The shift

AI drafts explanations at whatever complexity level you specify. You refine for accuracy and add the nuance AI misses.

The Basic Prompt

📋 Copy this template
Explain [concept] for [audience]. Target comprehension level: [beginner/intermediate/advanced] Length: [word count or paragraph count] Include: [what must be covered] Avoid: [jargon, technical terms, overly simple language]

Example:

Explain how HTTP requests work for junior web developers. Target comprehension level: Beginner (comfortable with HTML/CSS, new to backend) Length: 300-400 words Include: Request/response cycle, common methods (GET, POST), status codes Avoid: Deep technical implementation, network protocols beyond HTTP

What AI Does Well vs. What It Gets Wrong

✓ AI Strengths

  • Clear, structured explanations
  • Adjusts complexity on command
  • Multiple explanatory approaches
  • Examples and analogies, fast

✗ AI Limitations

  • Your learners' specific misconceptions
  • What prior knowledge they actually have
  • Industry-specific examples that resonate
  • Subtle technical accuracy in specialized fields

💡 The Division of Labor

AI creates the draft structure and language. You verify technical accuracy and ensure examples match your learners' context.

The Better Prompt

📋 Enhanced template with context
Explain [concept] for [specific audience description]. Context: - What they already know: [prerequisite knowledge] - What they struggle with: [common misconceptions] - Where they'll apply this: [performance environment] Structure: - Start with: [hook or motivating question] - Build from: [simple to complex / concrete to abstract] - Use: [analogies, examples, or diagrams] Constraints: - Length: [word count] - Reading level: [grade level or complexity] - Must include: [specific points] - Avoid: [terms, concepts, or approaches to skip]

Complexity Levels in Action

Same concept. Three audiences. Different everything.

📘 Beginner: Git for new developers ▼

"Git is like a save button with superpowers. Every time you 'commit' your code, Git creates a snapshot you can return to later. If you break something, you can jump back to when it worked. If your teammate changes the same file, Git helps you merge your changes together."

Notice: Uses everyday metaphors ("save button"), avoids technical terms

📗 Intermediate: Same topic, experienced devs ▼

"Git maintains a directed acyclic graph of commits, each representing the repository state at a point in time. When you commit, Git creates a hash-identified object containing your staged changes and metadata. Branches are simply movable pointers to commits, enabling parallel development workflows that merge via three-way algorithms."

Notice: Technical terminology (DAG, hashes), assumes workflow knowledge

📕 Advanced: Same topic, systems engineers ▼

"Git's object database stores four types: blobs (file contents), trees (directory structures), commits (repository snapshots with parent pointers), and tags (annotated references). The content-addressable storage model using SHA-1 hashes ensures data integrity while enabling distributed replication without coordination overhead."

Notice: Implementation details, assumes systems-level knowledge

The pattern: Same concept, different assumptions about what readers already know.

Quality Checklist

✓ Run every explanation through these checks ▼
Check Question
Right level Does it match what they actually know?
Accurate Is it technically correct? (AI makes mistakes)
Logical flow Does it build understanding step-by-step?
Concrete examples Are the examples relevant to their work?
Right length Not too vague (short) or overwhelming (long)?

🚩 Red Flag

AI sounds confident even when wrong. Always fact-check specialized content.

Examples and Analogies

AI generates examples fast. They're often generic.

❌ AI's generic example

"A database is like a filing cabinet. Each drawer is a table, and each folder is a row."

Abstract metaphor

✅ After you contextualize

"Think of your customer database like your company's CRM. Each customer record contains their name, email, purchase history, and support tickets. When you search for 'customers who bought Product X in Q4,' the database queries that table and returns matching rows—just like filtering your CRM by product and date."

Tools they actually use

Structural Patterns

Good explanations follow patterns. Tell AI which one to use:

📋 Four proven explanation structures ▼
Pattern When to use Prompt language
Problem → Solution Concept solves a specific pain "Start by describing the problem this solves, then explain how it works"
Simple → Complex Building layered understanding "Begin with the simplest version, then add complexity one layer at a time"
Concrete → Abstract Audience learns from examples "Start with a specific example, then generalize to the broader principle"
Compare/Contrast Similar concept already known "Explain by comparing it to [concept learners already know]"

Adjusting Complexity

Don't restart. Refine in place:

💬 Refinement prompts
Too complex? "Simplify this for someone with no technical background." Too simple? "Add more technical depth—assume the reader knows [prerequisite concept]." Wrong focus? "Rewrite to emphasize [specific aspect] rather than [current focus]." Missing examples? "Add a concrete example showing this in action in [specific context]."

Key Takeaways

  1. AI drafts structure and language. You ensure accuracy and context-fit.
  2. Specify complexity explicitly. Beginner/intermediate/advanced produces different content.
  3. Contextualize examples. Replace generic metaphors with tools learners recognize.
  4. Always fact-check. AI sounds confident even when wrong.

Try It Now

🎯 Your task:

Pick a concept from your current project. Generate explanations at three complexity levels. Check each for accuracy. Add industry-specific examples.

The test: Would an expert in your field spot any technical errors?

📥 Download: Explanation templates and quality checklist (PDF)

Ready-to-use templates with complexity calibration and fact-checking guides.

Download PDF