Hvordan-veiledninger burde være perfekte for AI-sitater, men våre blir ikke sitert. Hva er den hemmelige oppskriften?

Your hypothesis is exactly right. Let me explain why.

How AI processes how-to content:

When someone asks “How do I reset my password?”, AI looks for content it can extract as clear steps. It’s looking for:

1. Numbered or bulleted steps
2. Action verbs at the start (Click, Navigate, Enter)
3. Brief, complete instructions
4. Minimal surrounding text to parse through

Why comprehensive guides fail:

Your detailed explanations are great for humans who want to understand WHY. But AI is looking for WHAT to do. When step 1 is three paragraphs of context before the actual instruction, AI has trouble extracting it cleanly.

The solution: Layer your content.

Layer 1: Quick Answer (AI extracts this) Brief numbered steps, no fluff, direct instructions

Layer 2: Detailed Explanation (Humans who want more read this) Expandable sections or sub-pages with context

Example transformation:

Before: “First, you’ll want to navigate to the settings page. This is typically found in the top-right corner of your screen, represented by a gear icon. Once you’ve located this icon, click on it to open the settings menu…”

After: “Step 1: Click the gear icon in the top-right corner to open Settings.” [Expandable: Why this step matters / Common issues]

Same information. But the second version is extractable.

HowTo schema is critical for how-to guides.

Without schema:

• AI has to parse your HTML to understand steps
• May misinterpret structure
• Uncertain about what’s a step vs. what’s explanation

With HowTo schema:

• Steps are explicitly marked
• Each step has clear name, text, optional image
• AI systems can parse with confidence

Basic HowTo schema:

{
  "@type": "HowTo",
  "name": "How to Reset Your Password",
  "step": [
    {
      "@type": "HowToStep",
      "name": "Open Settings",
      "text": "Click the gear icon in the top-right corner."
    },
    {
      "@type": "HowToStep",
      "name": "Navigate to Security",
      "text": "Select 'Security' from the left menu."
    }
  ]
}

Our data:

How-to guides with HowTo schema: 42% AI citation rate How-to guides without: 18% AI citation rate

Same content quality. Schema makes the difference in AI understanding.

Practical template for AI-friendly how-to guides:

Title: “How to [Action] [Object]: [Brief Description]”

Quick Answer Section:

Quick Steps:

1. [Action verb] [what to do]
2. [Action verb] [what to do]
3. [Action verb] [what to do] Time required: X minutes What you’ll need: [Prerequisites]

Detailed Steps:

Step 1: [Action Title] [One-sentence instruction] [Why this step matters - optional] [Screenshot - optional] [Common issue for this step - optional]

[Repeat for each step]

Troubleshooting:

• Problem: X doesn’t work
• Solution: Try Y

FAQ:

• Q: Can I do this on mobile?
• A: Yes, the steps are similar…

Key elements:

• Action verbs starting each step
• Steps are standalone (make sense without reading previous step)
• Quick answer first, detail second
• Schema markup throughout

UX writing principles that help AI extraction:

1. Action Verbs First

• Bad: “The next thing you’ll do is click Submit”
• Good: “Click Submit”

2. One Action Per Step

• Bad: “Click Settings and then navigate to Privacy”
• Good: Two separate steps

3. Specific Over Vague

• Bad: “Go to the relevant section”
• Good: “Click the ‘Security’ tab”

4. Outcome Clarity

• Bad: “Enter your information”
• Good: “Enter your email address in the ‘Email’ field”

5. Conditional Clarity

• Bad: “You might need to verify”
• Good: “If prompted, click the verification link sent to your email”

These principles make content both user-friendly AND AI-extractable. Clear writing serves everyone.

Data on how-to guide characteristics and AI citations:

Combination effect:

Guides with ALL these characteristics: 67% citation rate Guides with NONE: 11% citation rate

What this means:

Each element helps. Combined, they’re powerful. Don’t just add schema - combine with concise steps, action verbs, and clear structure.

Quick note on images and video in how-to guides:

The problem: AI systems primarily parse text. An image showing “click here” doesn’t help AI understand the step.

The solution:

1. Every image needs descriptive alt text with the actual instruction
2. Critical information must be in text, not just visuals
3. Video transcripts should include step-by-step text

Example:

Bad: [Image of settings screen with arrow] Good: [Image of settings screen with arrow] Alt text: “Settings menu with Security option highlighted” Caption: “Click ‘Security’ in the left sidebar as shown above”

Images enhance for humans. Text enables AI understanding. Include both.

This thread solved my problem. Here’s my action plan:

Template Changes:

1. Add “Quick Steps” box at the very top
2. Keep detailed explanations but move below quick steps
3. Implement HowTo schema on all guides
4. Restructure steps: action verb first, one action per step, under 25 words

Immediate Priorities (Top 20 guides):

• Add quick steps section
• Add HowTo schema
• Review step clarity and brevity

Template for New Guides:

• Quick Answer section mandatory
• Detailed section follows
• Schema required before publish
• Checklist for action verbs, step length, prerequisites

Measurement:

• Track AI citations before/after with Am I Cited
• Compare restructured vs. old format

Key insight: Comprehensive is fine. But STRUCTURE matters more. Extractable quick steps + detailed explanation = best of both worlds.

Thanks everyone for the specific examples and data!