Learn what Help Center Audits are, how they help you maintain an effective knowledge base, and how to interpret key metrics and recommendations.
Help Center Audits are Fern's AI-powered analysis of your support conversations. Fern scans your connected support platforms (like Zendesk or Intercom) to identify documentation gaps, outdated content, and opportunities to reduce support volume through better self-service resources.
Why Help Center Audits Matter
Regular audits help you:
Identify Knowledge Gaps: Discover which topics frequently generate support tickets because they aren't adequately covered in your help center.
Reduce Support Volume: Create targeted articles for common questions so users can find answers themselves instead of contacting support.
Measure Content Effectiveness: Track whether your recent documentation updates are successfully reducing tickets on those topics.
Stay Current with User Needs: Understand seasonal patterns, product changes, and evolving customer questions.
Pro tip: Schedule weekly audits to catch documentation gaps early, before they turn into support ticket floods. Fern analyzes up to 3,000 of your most recent support conversations per audit.
How Audits Work
When you create an audit, Fern:
Imports up to 3,000 recent support conversations from your connected platforms
Reads and analyzes the conversations to understand customer questions
Identifies common questions by clustering similar requests together
Identifies content gaps by comparing customer questions to your existing help center articles
Generates prioritized recommendations for creating or updating articles
This process typically takes a few minutes, depending on your support volume. You'll receive an email notification when your audit is ready to review.
Key Metrics Explained
Each audit provides metrics that help you understand your help center's performance:
Conversations Analyzed: The total number of support conversations Fern reviewed during the audit period. This shows the volume of data used to generate insights.
Recommendations: The number of actionable suggestions Fern identified for improving your help center. Each recommendation represents an opportunity to create or update an article.
Coverage: The percentage of customer questions already addressed by existing articles in your help center. Higher coverage means your documentation is more comprehensive.
Estimated Impact: For each recommendation, Fern estimates the potential reduction in support tickets (e.g., "Get up to 15 fewer tickets every week"). Use this to prioritize which recommendations to implement first.
Understanding Recommendations
Fern organizes recommendations by priority level using a Fern Score (0-100), which factors in ticket frequency and escalation rates:
Top 5 Suggestions: The highest-priority recommendations that could significantly reduce your support volume.
Next 15 Suggestions: Medium-priority recommendations offering moderate impact.
Low Priority Suggestions: Other opportunities to review, typically addressing less frequent questions.
Each recommendation includes:
A clear title describing what content needs to be created or updated
A problem summary explaining the documentation gap
Supporting evidence from actual support tickets (expandable to view conversation details)
Actions to assign the work to Fern or reject the recommendation
When you assign a recommendation to Fern, the AI agent will draft the article for you. You can edit the prompt first to provide specific instructions, or let Fern work automatically.
Exploring Support Inbox Insights
Beyond recommendations, each audit includes a Support Inbox tab with additional metrics:
Total Intents: The number of unique question clusters identified across all conversations
Average Resolution Time: How long it typically takes to resolve support requests (e.g., "45 minutes")
Escalations: The breakdown of tickets escalated versus resolved (e.g., "20 escalated / 130 not")
Categories: A pie chart showing the distribution of request types (Questions, Bug Reports, Feature Requests, etc.)
Product Areas: A bar chart highlighting which parts of your product generate the most support requests (e.g., Billing, Integrations)
The Common Requests table shows clustered questions with integration sources and category indicators, helping you understand where users need the most help.
Audit Limits and Quotas
Understanding audit limits helps you plan your documentation strategy:
Free tril: Includes weekly audits analyzing up to 1,000 support conversations per month
Conversation Limit: Each audit analyzes up to 3,000 of your most recent support conversations
Audit Status Types
As you work with audits, you'll see different status badges:
Pending: Audit is queued and will start processing soon
In Progress: Fern is actively analyzing conversations and generating recommendations
Ready to Review: Audit is complete and recommendations are available
Reviewed: You've marked the audit as reviewed (this rejects all pending recommendations)
Failed: The audit encountered an error and couldn't complete
You can leave the audit detail page while processing is in progress. Fern will email you when your audit is ready to review, and you can return to it anytime from the Audits page.
What's Next
Now that you understand how Help Center Audits work, you can:
Create your first Help Center Audit to start identifying documentation gaps
Learn about connecting support platforms to power your audits
Set up scheduled audits to automatically monitor your help center on a regular cadence