Fern watches your support tickets and product changes to spot gaps in your documentation. Instead of waiting for patterns to become obvious, Fern surfaces them early so you can address issues before they pile up.
Where recommendations appear
All of Fern's suggestions show up in your Inbox. Think of it as your documentation to-do list, prioritized by what's actually affecting your users and what's changing in your product.
The Inbox organizes recommendations into three tabs:
Pending - Suggestions waiting for your review
Resolved - Recommendations you've acted on
Rejected - Suggestions you've dismissed
When you click into a recommendation, you'll see why Fern flagged it - specific tickets, commits, or releases that triggered the suggestion, along with what Fern thinks you should do about it.
How Fern identifies gaps
Fern monitors two types of signals that indicate missing or outdated documentation:
You can explore these raw signals in detail on the Signals dashboard.
Support patterns
When users reach out for help, it often means documentation is missing or unclear. Fern analyzes tickets from your support tools (Crisp, Intercom, Zendesk) looking for:
Questions that keep coming up repeatedly
Issues that take multiple back-and-forth messages to resolve
Topics tagged as high-priority or escalated
Searches that don't lead to helpful articles
The AI doesn't just count tickets - it understands context and intent to determine whether a documentation gap exists.
Product changes
When your product evolves, documentation needs to keep up. Fern tracks commits, pull requests, and releases from GitHub or Linear to catch:
New features that launched without accompanying docs
Breaking changes that will affect existing users
Updates to workflows or UI that make screenshots outdated
Deprecations that need migration guides
By connecting both support and codebase data, Fern can correlate product changes with user confusion - if a recent release starts generating support tickets, that's a strong signal.
The most actionable recommendations usually combine both signals - a product change that's already causing support volume.
How Fern prioritizes
Not all documentation gaps have equal impact. Fern assigns priority levels based on:
Volume - How many users are affected
Frequency - How often the issue comes up
Recency - Whether this is a new problem or ongoing
Impact - Whether it's blocking users or just nice-to-have
High-priority recommendations typically address widespread confusion or critical features without documentation. Low-priority suggestions might cover edge cases or minor improvements.
When recommendations are generated
Fern generates recommendations in two ways:
On-demand scans
You can trigger a scan whenever you want fresh insights. Choose a lookback period (7 days, 31 days, etc.) and Fern will analyze that window of data. Scans usually take 5-30 minutes depending on volume.
This is useful after major product launches, when you've just connected a new integration, or when you want to analyze a specific time period.
Automatic daily monitoring
Once you've set up recommendations, Fern runs daily scans in the background. This keeps your Inbox updated with emerging patterns without you having to remember to check.
Daily monitoring means you'll catch documentation gaps closer to when they happen, rather than discovering them weeks later when support volume has already spiked.
Scans analyze up to about 3,000 tickets. For higher volumes, Fern prioritizes the most recent and relevant data.
What Fern recommends
Recommendations come in a few different flavors:
Create new articles - When there's no documentation for something users are asking about
Update existing articles - When docs are outdated due to product changes
Expand coverage - When current documentation misses common scenarios
Each recommendation includes suggested actions - specific steps to address the gap. You can assign these directly to Fern as tasks, handle them yourself, or reject them if they're not relevant.
Acting on recommendations
From your Inbox, you have two main options:
Assign to Fern - This creates a task that Fern will draft for you. You'll review the draft before it goes live.
Reject - If a recommendation doesn't make sense or you've already addressed it, dismiss it. It won't come back.
The evidence panel shows you the actual tickets or commits behind each recommendation, so you can make informed decisions about what to prioritize.
Review evidence before acting - it gives you context about whether the recommendation is worth pursuing now or can wait.
What you need to get started
To receive recommendations, connect at least one support integration (Crisp, Intercom, or Zendesk). Adding a codebase integration (GitHub or Linear) is optional but gives Fern more complete visibility into documentation needs.
Once integrations are connected and you run your first scan, recommendations start flowing into your Inbox.