> ## Documentation Index
> Fetch the complete documentation index at: https://docs.buildbetter.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Taxonomy Overview

> Organize and categorize your signals with AI-powered product taxonomy

Taxonomy is BuildBetter's intelligent categorization system that helps you organize signals by product area. Using AI, BuildBetter generates a structured hierarchy based on your product documentation, then automatically labels incoming signals to keep everything organized.

## What is Taxonomy?

Taxonomy provides a four-level hierarchy for categorizing your customer feedback and signals:

<CardGroup cols={2}>
  <Card title="Domain" icon="building">
    High-level product areas or customer journey stages (e.g., "Onboarding", "Core Platform", "Enterprise")
  </Card>

  <Card title="Product" icon="box">
    Specific offerings, modules, or major feature sets within a domain (e.g., "Recording Service", "Analytics Dashboard")
  </Card>

  <Card title="Feature" icon="gear">
    Individual capabilities or functionality (e.g., "Speaker Assignment", "Export Options", "Search Filters")
  </Card>

  <Card title="Tag" icon="tag">
    Specific instances, variations, or fine-grained labels (e.g., "CSV Export", "PDF Export", "API Export")
  </Card>
</CardGroup>

## Why Use Taxonomy?

<Tabs>
  <Tab title="Organization">
    **Keep signals organized at scale**

    As your team captures thousands of signals from calls, support tickets, and feedback forms, taxonomy provides structure:

    * Group related feedback together automatically
    * Find all signals related to a specific feature instantly
    * Track feedback trends by product area over time
  </Tab>

  <Tab title="AI Auto-Labeling">
    **Let AI do the categorization**

    Once your taxonomy is set up, BuildBetter automatically labels new signals:

    * No manual tagging required for routine categorization
    * Consistent labeling across your entire team
    * Backfill existing signals with new taxonomy labels
  </Tab>

  <Tab title="Analysis">
    **Enable deeper insights**

    Taxonomy powers better analysis and reporting:

    * Filter signals by product area in seconds
    * Generate reports scoped to specific features
    * Identify which product areas get the most feedback
    * Compare sentiment across different domains
  </Tab>
</Tabs>

## How Taxonomy Works

### The Hierarchy Structure

```
Domain (Level 1)
├── Product (Level 2)
│   ├── Feature (Level 3)
│   │   ├── Tag (Level 4)
│   │   └── Tag
│   └── Feature
│       └── Tag
└── Product
    └── Feature
        └── Tag
```

<Note>
  **Auto-labeling levels**: Product, Feature, and Tag levels are automatically applied to signals by AI. Domain is intentionally not auto-labeled to encourage high-level human decision-making about product organization.
</Note>

### Example Taxonomy

Here's what a taxonomy might look like for a project management tool:

| Domain             | Product       | Feature         | Tag          |
| ------------------ | ------------- | --------------- | ------------ |
| Project Management | Task Board    | Kanban View     | Drag & Drop  |
| Project Management | Task Board    | Kanban View     | Swimlanes    |
| Project Management | Task Board    | List View       | Sorting      |
| Project Management | Time Tracking | Timer           | Manual Entry |
| Collaboration      | Comments      | Mentions        | @User        |
| Collaboration      | Comments      | Threads         | Reply        |
| Integrations       | Calendar Sync | Google Calendar | Two-way Sync |
| Integrations       | Calendar Sync | Outlook         | Import Only  |

## Getting Started with Taxonomy

<Steps>
  <Step title="Generate Your Taxonomy">
    Navigate to **Settings > Features > AI Labeling** and click **Generate Taxonomy** in the Custom Taxonomy section. You'll be prompted to provide product documentation or a description of your product.

    <Tip>
      The more detailed your input, the better your taxonomy will be. Include feature lists, documentation pages, or a comprehensive product description.
    </Tip>
  </Step>

  <Step title="Review and Refine">
    After AI generates your taxonomy, review the hierarchy in the Taxonomy Editor. You can:

    * Rename nodes to match your team's terminology
    * Add missing features or products
    * Remove irrelevant categories
    * Reorganize the hierarchy structure
  </Step>

  <Step title="Enable Auto-Labeling">
    Once your taxonomy is finalized, BuildBetter will automatically apply relevant labels to new signals. You can also backfill existing signals with the new taxonomy.
  </Step>
</Steps>

## The Taxonomy Editor

The Taxonomy Editor provides full control over your product hierarchy:

### Viewing Your Taxonomy

<AccordionGroup>
  <Accordion title="Tree View" icon="sitemap">
    See your entire taxonomy as an expandable tree structure. Click the arrow next to any node to expand or collapse its children.
  </Accordion>

  <Accordion title="Level Indicators" icon="layer-group">
    Each level is visually distinguished:

    * **Domain**: Top-level organizational categories
    * **Product**: Major offerings (auto-labelable)
    * **Feature**: Specific capabilities (auto-labelable)
    * **Tag**: Fine-grained labels (auto-labelable)
  </Accordion>

  <Accordion title="Signal Counts" icon="chart-bar">
    Each taxonomy node displays a count of how many signals have been labeled with that category. This helps you:

    * Identify which product areas receive the most feedback
    * Spot under-represented features that may need more customer attention
    * Track how feedback is distributed across your product
    * Understand the relative importance of different features based on customer mentions
  </Accordion>
</AccordionGroup>

### Editing Operations

| Action               | Description                                                     |
| -------------------- | --------------------------------------------------------------- |
| **Add Node**         | Click the + button on any node to add a child at the next level |
| **Edit Name**        | Click on any node name to edit it inline                        |
| **Edit Description** | Add context to help AI understand when to apply this label      |
| **Delete Node**      | Remove a node and all its children                              |
| **Reorder**          | Drag nodes to change their order within a level                 |

### AI Labeling Instructions

For each node, you can provide custom instructions to guide the AI when auto-labeling:

```
Example for "Mobile App" feature:
"Apply this label when the customer mentions the iOS app,
Android app, mobile experience, or phone-based usage."
```

<Info>
  Custom instructions are especially useful for:

  * Ambiguous features that might overlap
  * Product-specific terminology your customers use
  * Features with multiple common names
</Info>

## Auto-Labeling Signals

### How Auto-Labeling Works

When a new signal is extracted:

1. **Content Analysis**: AI reads the signal's summary and context
2. **Taxonomy Matching**: Compares content against your taxonomy definitions
3. **Label Application**: Applies the most relevant Product, Feature, and Tag labels
4. **Confidence Check**: Only applies labels when confident about the match

### Managing Auto-Fill Jobs

You can trigger bulk auto-labeling for existing signals:

<Tabs>
  <Tab title="Backfill Mode">
    Only labels signals that don't have taxonomy values yet. Use this when adding taxonomy to an existing workspace.
  </Tab>

  <Tab title="Overwrite Mode">
    Re-labels all signals, replacing existing taxonomy values. Use this after significant taxonomy changes.
  </Tab>
</Tabs>

<Warning>
  Auto-labeling uses credits based on the number of signals processed. Check your usage before running large backfill jobs.
</Warning>

## Using Taxonomy with Signals

### Filtering by Taxonomy

On the Signals page, use taxonomy filters to narrow down results:

* **Domain Filter**: See all signals in a product area
* **Product Filter**: Focus on a specific offering
* **Feature Filter**: Find feedback about particular functionality
* **Tag Filter**: Get granular with specific variations

### Taxonomy in Slack

When creating signals from Slack messages, you can select taxonomy labels:

1. Use the BuildBetter Slack app to create a signal
2. In the signal creation modal, select relevant taxonomy tags
3. The signal is created with your selected categorization

<Tip>
  Selecting taxonomy labels during Slack signal creation overrides auto-labeling for that signal, giving you manual control when needed.
</Tip>

## Best Practices

<Check>
  **Start broad, then refine**: Begin with major product areas and add detail as you learn what categorizations are most useful.
</Check>

<Check>
  **Use your customers' language**: Name taxonomy nodes using terms your customers actually use in conversations.
</Check>

<Check>
  **Add descriptions**: Help the AI by adding clear descriptions to ambiguous or similar-sounding features.
</Check>

<Check>
  **Review periodically**: As your product evolves, update your taxonomy to reflect new features and retired ones.
</Check>

<Check>
  **Don't over-categorize**: A taxonomy that's too granular creates noise. Focus on distinctions that matter for analysis.
</Check>

## Taxonomy vs. Tags

| Taxonomy                | Custom Tags        |
| ----------------------- | ------------------ |
| Hierarchical (4 levels) | Flat list          |
| AI-generated structure  | Manually created   |
| Auto-applied to signals | Manually applied   |
| Product-focused         | Flexible use cases |
| System-managed          | User-managed       |

<Info>
  Use taxonomy for product categorization and custom tags for cross-cutting concerns like "urgent", "competitor mention", or "follow-up needed".
</Info>

## FAQ

<AccordionGroup>
  <Accordion title="Can I have multiple taxonomies?">
    Currently, BuildBetter supports one product taxonomy per workspace. This ensures consistent categorization across all signals.
  </Accordion>

  <Accordion title="What happens if I delete a taxonomy node?">
    Deleting a node removes it from the hierarchy. Existing signals that were labeled with that node will retain the label as historical data, but new signals won't be categorized with the deleted node.
  </Accordion>

  <Accordion title="How accurate is auto-labeling?">
    Auto-labeling accuracy depends on your taxonomy quality and the clarity of your signal content. Well-described taxonomy nodes with clear instructions achieve higher accuracy. You can always manually adjust labels if needed.
  </Accordion>

  <Accordion title="Can I regenerate my taxonomy?">
    Yes, you can regenerate your taxonomy at any time. This will replace your existing hierarchy, so export or note any customizations you want to preserve before regenerating.
  </Accordion>
</AccordionGroup>

Taxonomy transforms your signal library from a collection of feedback into a structured knowledge base organized around your product. Combined with AI auto-labeling, it ensures every piece of customer feedback finds its proper place automatically.