> ## 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.

# Tag Groups Overview

> Create intelligent tag collections with AI-powered auto-labeling rules

Tag Groups let you organize related tags together and define AI-powered rules for automatically applying them to your calls and signals. Instead of manually tagging every piece of feedback, set up your rules once and let BuildBetter handle the classification.

## What Are Tag Groups?

A Tag Group is a collection of related tags with shared labeling instructions. Each group can:

<CardGroup cols={2}>
  <Card title="Group Related Tags" icon="layer-group">
    Organize tags that belong together (e.g., "Customer Segments", "Product Areas", "Feedback Types")
  </Card>

  <Card title="Define AI Rules" icon="robot">
    Provide instructions that tell AI when to apply each tag in the group
  </Card>

  <Card title="Auto-Label Content" icon="wand-magic-sparkles">
    Automatically tag calls and signals based on your defined rules
  </Card>

  <Card title="Process Retroactively" icon="clock-rotate-left">
    Apply new tag rules to existing content with one click
  </Card>
</CardGroup>

## Tag Groups vs. Individual Tags

| Individual Tags           | Tag Groups                         |
| ------------------------- | ---------------------------------- |
| Applied manually          | Applied automatically by AI        |
| No classification rules   | Includes AI labeling instructions  |
| Standalone labels         | Organized into logical collections |
| One-at-a-time application | Bulk auto-tagging                  |

<Tip>
  Use individual tags for ad-hoc labeling during review. Use tag groups when you want consistent, automated classification across all your content.
</Tip>

## How Tag Groups Work

<Steps>
  <Step title="Create a Tag Group">
    Define a group with a name and optional high-level instructions explaining what this group is for.

    **Example**: "Customer Segments" with instructions: "Classify based on company size and industry mentioned in the conversation."
  </Step>

  <Step title="Add Tags as Members">
    Add existing tags to the group. For each tag, you can provide specific instructions for when it should be applied.

    **Example Tags**:

    * "Enterprise" - Apply when company has 1000+ employees or mentions enterprise features
    * "SMB" - Apply for small to medium businesses under 500 employees
    * "Startup" - Apply when company mentions being early-stage or recently founded
  </Step>

  <Step title="Configure Targets">
    Choose what content types this tag group should automatically label:

    * **Calls/Interviews**: Tag your recorded meetings
    * **Signals/Extractions**: Tag extracted insights from conversations
  </Step>

  <Step title="AI Applies Tags">
    When new content is processed, AI reads the transcript or signal content and applies appropriate tags from the group based on your instructions.
  </Step>
</Steps>

## Creating a Tag Group

Navigate to **Settings > Tag Groups** to manage your tag groups.

### Basic Setup

1. Click **Create Tag Group**
2. Enter a group name (e.g., "Feedback Categories")
3. Add optional group-level instructions

<Info>
  Group-level instructions provide context that applies to all tags in the group. Use them to explain the overall classification goal.
</Info>

### Adding Members

After creating the group, add tags as members:

1. Click **Add Tags** on your tag group
2. Select existing tags or create new ones
3. For each tag, optionally add specific instructions

**Example Instructions for a "Priority Level" Tag Group:**

| Tag      | Instructions                                                              |
| -------- | ------------------------------------------------------------------------- |
| Critical | Apply when customer mentions deal-breaker, blocker, or threatens to churn |
| High     | Apply for significant pain points affecting daily workflows               |
| Medium   | Apply for improvements that would be nice to have                         |
| Low      | Apply for minor suggestions or cosmetic feedback                          |

### Tag-Level Instructions

Per-tag instructions help AI make precise decisions:

```
Tag: "Integration Request"
Instructions: "Apply when the customer asks about connecting
BuildBetter with another tool they use, mentions API access,
or discusses data sync between systems."
```

<Tip>
  Be specific in your instructions. Include example phrases, keywords, or scenarios that should trigger each tag.
</Tip>

## Tag Group Settings

### Multi-Tag Selection

<Tabs>
  <Tab title="Single Tag Mode">
    Only one tag from the group can be applied per item. Use this for mutually exclusive categories like:

    * Customer segments (Enterprise vs SMB vs Startup)
    * Priority levels (Critical vs High vs Medium vs Low)
    * Deal stages (Discovery vs Evaluation vs Negotiation)
  </Tab>

  <Tab title="Multi-Tag Mode">
    Multiple tags from the group can be applied to the same item. Use this for non-exclusive categories like:

    * Product areas mentioned (can discuss multiple features)
    * Departments involved (Sales AND Success on the same call)
    * Topics covered (Pricing AND Roadmap AND Competition)
  </Tab>
</Tabs>

### Skip Tagging Option

* **Required tagging**: Every item must receive at least one tag from this group
* **Optional tagging**: Items can have no tags if none apply

<Note>
  Enable skip tagging for groups where not every conversation will match. For example, a "Competitor Mentions" group should allow skipping when no competitors are discussed.
</Note>

## Targeting Content Types

Tag groups can target different content types:

### Calls & Interviews

Tags are applied to the overall call/meeting record based on transcript analysis.

**Best for:**

* Meeting types (Demo, Discovery, Support)
* Participants (Prospect, Customer, Partner)
* Outcomes (Successful, Needs Follow-up, At Risk)

### Signals & Extractions

Tags are applied to individual extracted insights from conversations.

**Best for:**

* Feedback categories (Feature Request, Bug Report, Praise)
* Sentiment indicators (Frustrated, Satisfied, Neutral)
* Action items (Follow-up Required, Escalate, No Action)

## Retroactive Processing

When you create or update a tag group, you can apply it to existing content:

1. Go to your tag group settings
2. Enable **Retroactive Processing**
3. BuildBetter will queue a job to process historical content

<Warning>
  Retroactive processing uses credits based on the number of items processed. Check your usage before processing large datasets.
</Warning>

## Viewing Tag Group Results

### On Calls

Tagged calls show their applied tags in the call detail view and can be filtered in the calls list.

### On Signals

Filter signals by tag group tags to find specific categories of feedback.

### Tag Group Logs

BuildBetter logs all auto-tagging decisions for transparency:

* What content was analyzed
* Which tags were considered
* Why specific tags were applied
* AI confidence in the decision

## Best Practices

<Check>
  **Start with clear categories**: Define tag groups around distinct classification needs (who, what, why, priority).
</Check>

<Check>
  **Write specific instructions**: Include keywords, phrases, and examples that should trigger each tag.
</Check>

<Check>
  **Use single-tag mode for exclusive categories**: Prevent conflicting tags like "Enterprise" AND "Startup" on the same call.
</Check>

<Check>
  **Enable skip tagging when appropriate**: Not every conversation will match every tag group.
</Check>

<Check>
  **Review and refine**: Check tag group logs periodically to see if AI is making correct decisions. Adjust instructions as needed.
</Check>

<Check>
  **Combine with taxonomy**: Use tag groups for cross-cutting concerns (priority, sentiment) and taxonomy for product categorization.
</Check>

## Example Tag Groups

### Feedback Type

| Tag             | Instructions                                         |
| --------------- | ---------------------------------------------------- |
| Feature Request | Customer asks for new functionality or capabilities  |
| Bug Report      | Customer reports something not working as expected   |
| Usability Issue | Customer struggles with how something works          |
| Praise          | Customer expresses satisfaction or gives compliments |
| Question        | Customer asks how to do something                    |

### Deal Risk Level

| Tag                 | Instructions                                              |
| ------------------- | --------------------------------------------------------- |
| At Risk             | Mentions competitors, budget concerns, or timeline delays |
| On Track            | Positive signals, engaged stakeholders, clear next steps  |
| Champion Identified | Strong internal advocate pushing for purchase             |
| Blocked             | Specific blocker preventing progress                      |

### Customer Lifecycle

| Tag        | Instructions                            |
| ---------- | --------------------------------------- |
| Prospect   | Pre-purchase evaluation phase           |
| Onboarding | Recently signed, getting started        |
| Active     | Regular usage, established customer     |
| Expansion  | Discussing additional seats or features |
| At Risk    | Churn signals, reduced engagement       |

## FAQ

<AccordionGroup>
  <Accordion title="How many tag groups can I create?">
    There's no limit to the number of tag groups. Create as many as needed for your classification needs.
  </Accordion>

  <Accordion title="Can a tag belong to multiple groups?">
    Yes, the same tag can be a member of multiple tag groups with different instructions in each context.
  </Accordion>

  <Accordion title="What happens if tag group instructions conflict?">
    Each tag group is evaluated independently. If the same tag is in multiple groups, it may be applied based on different criteria from each group.
  </Accordion>

  <Accordion title="How accurate is auto-tagging?">
    Accuracy depends on the clarity of your instructions. Specific, detailed instructions with examples produce better results than vague descriptions.
  </Accordion>

  <Accordion title="Can I override auto-applied tags?">
    Yes, you can manually add or remove tags from any call or signal, regardless of how they were applied.
  </Accordion>
</AccordionGroup>

Tag Groups transform manual tagging into intelligent, automated classification. Define your rules once, and let AI consistently apply them across all your customer conversations.