Audience Analysis & Documentation Structure

🔄 Quick Recall: In the previous lesson, you learned why documentation matters and how AI transforms technical writing. Now you’ll learn the foundation of all good documentation: understanding who you’re writing for and choosing the right structure to serve them.

Before writing a single word, two decisions shape everything: who is the reader and what do they need? Get these wrong, and even beautifully written documentation fails. AI helps by analyzing your existing docs, identifying audience gaps, and structuring content according to proven frameworks.

The Diátaxis Framework

AI prompt for documentation audit:

Audit my existing documentation using the Diátaxis framework. Documentation: [PASTE OR DESCRIBE YOUR CURRENT DOCS]. Categorize each page as: (1) TUTORIAL — learning-oriented, step-by-step, hands-on (for newcomers), (2) HOW-TO — task-oriented, goal-focused, assumes knowledge (for practitioners), (3) REFERENCE — information-oriented, exhaustive, structured for lookup (for experienced users), (4) EXPLANATION — understanding-oriented, provides context and rationale (for curious minds). Identify: which quadrants are well-covered, which are missing, and specific recommendations for filling gaps. Most documentation is heavy on reference and light on tutorials — confirm or refute this for my docs.

The four documentation types:

Audience Analysis

AI prompt for audience profiling:

Create audience profiles for my documentation. Product: [DESCRIBE]. Users include: [LIST USER TYPES — developers, admins, end users, etc.]. For each audience: (1) their technical level (beginner, intermediate, expert), (2) what they’re trying to accomplish (tasks, not features), (3) their preferred documentation format (quick start, video, reference, example code), (4) the terminology they use (which might differ from our internal terms), (5) the documentation entry point they’ll likely use (Google search, in-app help, README). Generate: audience-specific content strategies and the documentation types each audience needs most.

✅ Quick Check: You write “Configure the webhook endpoint by specifying the callback URL in the integration settings.” A non-technical product manager reads this. What’s wrong? (Answer: Three terms the PM might not understand — webhook, endpoint, callback URL. For a technical audience, this is fine. For a PM, rewrite: “Set up automatic notifications by entering the web address where you want to receive updates.” Same information, different words. AI rewrites content for different audiences automatically while preserving accuracy.)

Information Architecture

AI prompt for doc site structure:

Design the information architecture for my documentation site. Product: [DESCRIBE]. User types: [LIST]. Current pain points: [WHAT USERS CAN’T FIND]. Generate: (1) top-level navigation structure (max 6-8 categories), (2) page hierarchy within each category, (3) cross-references between related pages, (4) recommended search keywords for each page (what users type when looking for this information), (5) landing page content for each section. The structure should match how users think about tasks, not how the product is organized internally.

Progressive Disclosure

AI prompt for content layering:

Restructure this documentation using progressive disclosure. Current content: [PASTE CONTENT]. Create three layers: (1) Layer 1 (Quick Start) — the minimum needed to accomplish the most common task. Should be completable in under 5 minutes. (2) Layer 2 (Details) — configuration options, error handling, edge cases. For readers who need more than the happy path. (3) Layer 3 (Advanced) — internals, troubleshooting rare issues, performance tuning. For power users only. Each layer should be self-contained — a reader at Layer 1 should never feel like they’re missing something essential.

Key Takeaways

• The Diátaxis framework (tutorials, how-tos, reference, explanation) solves the most common documentation problem: mixing learning content with reference content. Most doc sets are heavy on reference and light on tutorials — AI audits your docs and identifies the gaps
• Audience analysis before writing prevents the single biggest documentation failure: writing for yourself instead of the reader. AI generates audience profiles and rewrites content for different technical levels automatically
• Progressive disclosure matches content depth to reader needs: 80% of readers need Layer 1 (quick start), 15% need Layer 2 (details), 5% need Layer 3 (advanced) — structure your docs so the most-needed content is the easiest to find
• Documentation structure should match how users think about tasks, not how the product is organized internally — “How do I reset my password” vs. “Authentication module > Password management”
• AI adapts the same content for multiple audiences: describe a feature once, and AI generates developer-focused, PM-focused, and end-user-focused versions with appropriate terminology and detail level

Up Next

In the next lesson, you’ll create API documentation with AI — endpoint references, authentication guides, code examples, and the documentation that makes APIs easy to adopt.