> ## 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.

# Clustering & Insights

> AI-powered theme detection and visual analytics for customer feedback

<Note>
  **About "Clustering"**: In BuildBetter, "Clustering" refers to both the AI-powered thematic grouping of signals AND the dashboard/visualization system. When you see "Clustering" or "Clusters" in the navigation, it's where you'll find signal analysis dashboards, visualizations, and AI-generated theme reports.
</Note>

BuildBetter's Clustering feature uses advanced AI to automatically group similar customer feedback into themes, revealing patterns and trends you might otherwise miss. It also provides powerful dashboards and visualizations to explore your signal data. Transform thousands of individual signals into actionable insights with AI analysis and visual analytics.

## Understanding Clustering

Clustering automatically organizes your customer feedback into meaningful themes:

<CardGroup cols={2}>
  <Card title="AI-Powered Grouping" icon="brain">
    Machine learning identifies similar feedback and groups them by semantic meaning, not just keywords
  </Card>

  <Card title="Dynamic Analysis" icon="sync">
    Clusters update in real-time as new feedback arrives, keeping insights current
  </Card>

  <Card title="Trend Detection" icon="chart-line">
    Automatically identifies rising and declining themes to spot emerging issues early
  </Card>

  <Card title="Visual Analytics" icon="chart-pie">
    Rich dashboards with 15+ visualization types for different analytical perspectives
  </Card>
</CardGroup>

## How Clustering Works

<Steps>
  <Step title="Automatic Processing">
    As signals are extracted from calls, chats, and feedback, the AI analyzes their semantic content
  </Step>

  <Step title="Theme Identification">
    Similar feedback is grouped into clusters based on meaning, creating themes like "onboarding issues" or "pricing concerns"
  </Step>

  <Step title="Continuous Refinement">
    Clusters evolve as more data arrives, becoming more accurate and nuanced over time
  </Step>

  <Step title="Insight Generation">
    AI generates reports for each cluster with key insights, root causes, and recommendations
  </Step>
</Steps>

<Note>
  Clusters require a minimum of 3 similar signals to form, ensuring themes represent actual patterns rather than one-off feedback.
</Note>

## Navigating Clusters

### Cluster List View

Access your clusters through the main clustering page:

<Tabs>
  <Tab title="All Clusters">
    Browse all identified themes with:

    * Cluster names and descriptions
    * Signal counts and growth indicators
    * Trending badges for rising themes
    * Search and filter capabilities
  </Tab>

  <Tab title="Trending Clusters">
    Focus on clusters with positive growth:

    * Percentage increase indicators
    * Time period comparisons
    * Quick access to emerging issues
    * Priority sorting by growth rate
  </Tab>

  <Tab title="Favorites">
    Star important clusters for quick access:

    * Monitor critical themes
    * Track resolution progress
    * Share with team members
    * Create focused dashboards
  </Tab>
</Tabs>

### Cluster Detail Pages

Click any cluster to explore in-depth:

<AccordionGroup>
  <Accordion title="AI-Generated Reports" icon="file-lines">
    Each cluster includes timestamped reports with:

    * **Key Insights**: Main patterns identified
    * **Root Cause Analysis**: Underlying issues
    * **Business Impact**: How this affects your company
    * **Recommendations**: Suggested actions
    * **Trend Analysis**: Growth or decline patterns
  </Accordion>

  <Accordion title="Customer Quotes" icon="quote">
    Representative feedback from real customers:

    * Direct quotes with context
    * Source attribution with timestamps
    * Click-through to original recordings
    * Sentiment indicators
  </Accordion>

  <Accordion title="Impact Analysis" icon="chart-simple">
    Understand who and what is affected:

    * Company breakdown
    * User persona distribution
    * Revenue impact potential
    * Product area mapping
  </Accordion>
</AccordionGroup>

## Creating Custom Dashboards

Build visual analytics dashboards tailored to your needs. Access this from the **Clustering** section in the main navigation.

<Note>
  **Finding Dashboards**: Navigate to Signals or Clustering in BuildBetter to access dashboard creation and visualization tools.
</Note>

### Dashboard Builder

<Steps>
  <Step title="Start Building">
    Go to Clustering section and click "Customize Dashboard" to enter edit mode
  </Step>

  <Step title="Add Visualizations">
    Choose from 15+ card types and drag them onto your dashboard
  </Step>

  <Step title="Configure Each Card">
    * Select data source and filters
    * Choose visualization style
    * Set card width (full, half, third, quarter)
    * Add titles and descriptions
  </Step>

  <Step title="Arrange Layout">
    Drag and drop cards to create your ideal layout
  </Step>

  <Step title="Save and Share">
    Save your dashboard and share the URL with teammates
  </Step>
</Steps>

### Trends Category

The card picker includes a dedicated **Trends** category for time-based views. Use these cards when you want to see whether a cluster, saved search, or filtered signal set is changing over time rather than just reading a point-in-time total.

| Trend Card              | What It Shows                       | Best For                                                                    |
| ----------------------- | ----------------------------------- | --------------------------------------------------------------------------- |
| **Volume Trend**        | Signal count over time              | Spotting rising themes or declining issue volume                            |
| **Sentiment Trend**     | Average sentiment over time         | Seeing whether feedback is getting more positive or negative                |
| **Severity Trend**      | Average severity over time          | Monitoring whether issues are becoming more urgent                          |
| **Impact Distribution** | Impact mix over the selected period | Comparing revenue, adoption, satisfaction, retention, and efficiency impact |

Each trend card supports **7, 30, 90, and 365 day** ranges. The selected range is saved per card, so one dashboard can show a 7-day operational view next to a 90-day planning view.

### Creating Dashboards from Taxonomy

Quickly create pre-configured dashboards based on your product taxonomy. This is the fastest way to set up focused analytics for specific product areas.

<Steps>
  <Step title="Access Templates">
    Click "New Clustering" and select "Use Clustering Template"
  </Step>

  <Step title="Choose Source">
    Select the "From Taxonomy" tab to see your product hierarchy
  </Step>

  <Step title="Select Nodes">
    Browse your taxonomy tree and check the product areas you want to analyze. You can select multiple nodes to create multiple dashboards at once.
  </Step>

  <Step title="Create Dashboards">
    Click "Create Clustering" to generate dashboards. Each selected node creates a separate dashboard pre-configured with:

    * Filters scoped to that taxonomy node
    * External customer feedback only (by default)
    * Last 30 days of data
    * Standard visualization cards
  </Step>
</Steps>

**Default Dashboard Cards:**

* Total Signals count
* Total Clusters count
* Average Sentiment score
* Average Severity score
* Top Clusters by Signal Count (bar chart)
* Customer Voice by Cluster (quotes)
* Cluster Details

<Tip>
  Creating dashboards from taxonomy is ideal for product managers who want to quickly set up monitoring for their product areas without manual configuration.
</Tip>

### Dashboard Templates

In addition to taxonomy-based creation, you can also use pre-built templates for common use cases:

| Template          | Description                                |
| ----------------- | ------------------------------------------ |
| Feature Requests  | Track what customers are asking for        |
| Churn Issues      | Monitor signals indicating potential churn |
| Customer Feedback | General customer sentiment and feedback    |
| Customer Issues   | Track reported problems and complaints     |
| Adoption Issues   | Monitor onboarding and adoption challenges |
| Competitors       | Track competitive mentions                 |
| Testimonials      | Collect customer success stories           |
| Compliments       | Positive feedback and praise               |

### Available Visualizations

<CardGroup cols={2}>
  <Card title="Bar Charts" icon="chart-column">
    * Top clusters by volume
    * Cluster growth comparisons
    * Category breakdowns
    * Time period analysis
  </Card>

  <Card title="Pie Charts" icon="chart-pie">
    * Persona distribution
    * Sentiment breakdown
    * Category allocation
    * Company distribution
  </Card>

  <Card title="Ridge Charts" icon="mountain">
    * Sentiment distribution curves
    * Severity patterns
    * Bias analysis
    * Emotion distributions
  </Card>

  <Card title="Heatmaps" icon="border-all">
    * Cluster activity over time
    * Cross-category correlations
    * Temporal patterns
    * Intensity mapping
  </Card>
</CardGroup>

### Advanced Visualizations

<Tabs>
  <Tab title="Timeline Charts">
    Track cluster evolution:

    * Signal volume over time
    * Growth rate trends
    * Seasonal patterns
    * Event correlations
  </Tab>

  <Tab title="Statistical Charts">
    Deep analytical views:

    * Violin plots for distributions
    * Box plots for ranges
    * Scatter plots for correlations
    * Regression analysis
  </Tab>

  <Tab title="Metric Cards">
    Key performance indicators:

    * Total signal counts
    * Average sentiment scores
    * Growth percentages
    * Impact metrics
  </Tab>
</Tabs>

## Quick Filters

Speed up your analysis with pre-built filter templates that apply common filter combinations with one click.

### Available Quick Filters

<Tabs>
  <Tab title="Boundary Filters">
    * **External Only**: Signals from customers and external users
    * **Internal Only**: Signals from your internal team
  </Tab>

  <Tab title="External Feedback">
    * **Feature Requests**: What customers are asking for
    * **Customer Complaints**: Issues raised by customers
    * **High Severity**: Critical customer issues
    * **Urgent Issues**: High severity combined with negative sentiment
    * **Testimonials**: Customer success stories
    * **Compliments**: Explicit positive mentions
  </Tab>

  <Tab title="Impact Analysis">
    * **Churn Risk**: Signals indicating potential churn
    * **Adoption Issues**: Signals about product adoption challenges
    * **Expansion Opportunities**: Signals indicating growth potential
    * **Low Bias**: Unbiased, objective feedback
  </Tab>

  <Tab title="Internal">
    * **Internal Blockers**: Issues raised by your team
  </Tab>
</Tabs>

### Using Quick Filters

Quick filters appear as clickable buttons above your signal list or clustering dashboard. Click any button to instantly apply that filter combination to your current view.

<Tip>
  Quick filters can be combined with additional manual filters using the query builder for even more specific analysis.
</Tip>

## Filtering and Analysis

### Query Builder

Create sophisticated filters using the visual query builder:

<AccordionGroup>
  <Accordion title="Natural Language Queries" icon="message">
    Type queries in plain English:

    * "Show negative feedback about onboarding from last month"
    * "Find pricing concerns from enterprise customers"
    * "What are new users complaining about?"

    The AI converts these to structured filters automatically.
  </Accordion>

  <Accordion title="Advanced Filters" icon="filter">
    Build complex queries with:

    * AND/OR logic combinations
    * Date range selections
    * Sentiment thresholds
    * Company/persona filters
    * Custom property matching
  </Accordion>

  <Accordion title="Saved Filters" icon="save">
    Save frequently used filter combinations:

    * Name your filter sets
    * Apply with one click
    * Share with teammates
    * Use across dashboards
  </Accordion>
</AccordionGroup>

## Cluster Reports

### Understanding AI Reports

Each cluster generates comprehensive reports containing:

**Trend Analysis**

* Current period volume vs. previous
* Percentage change calculations
* Growth trajectory predictions
* Seasonal pattern identification

**Key Insights**

* Main themes identified
* Common pain points
* Satisfaction drivers
* Emerging concerns

**Root Cause Analysis**

* Underlying issues identified
* Contributing factors
* System dependencies
* Process breakdowns

**Business Impact**

* Revenue implications
* Churn risk assessment
* Adoption barriers
* Competitive threats

### Using Reports Effectively

<Steps>
  <Step title="Regular Review">
    Check trending clusters weekly to catch emerging issues early
  </Step>

  <Step title="Share Insights">
    Copy report content or share links with stakeholders
  </Step>

  <Step title="Track Progress">
    Monitor how clusters evolve after taking action
  </Step>

  <Step title="Create Actions">
    Use recommendations to drive product and process improvements
  </Step>
</Steps>

## Best Practices

<Check>
  **Monitor trending clusters weekly**: Early detection of emerging issues prevents escalation
</Check>

<Check>
  **Create role-specific dashboards**: Product, support, and sales teams need different views
</Check>

<Check>
  **Combine filters thoughtfully**: Layer time, sentiment, and persona filters for deeper insights
</Check>

<Check>
  **Act on insights**: Use cluster reports to drive concrete improvements
</Check>

<Check>
  **Share visualizations**: Export dashboards for presentations and team alignment
</Check>

## Common Use Cases

### Product Management

* Track feature request themes
* Identify usability issues
* Monitor adoption barriers
* Prioritize roadmap items

### Customer Success

* Spot at-risk accounts early
* Identify common support themes
* Track satisfaction trends
* Prevent churn patterns

### Marketing & Sales

* Understand objections
* Track competitive mentions
* Identify value propositions
* Monitor market trends

### Leadership

* Strategic theme tracking
* Quarterly trend analysis
* Department performance
* Market intelligence

<Warning>
  **Data Requirements**: Clustering works best with 100+ signals. Smaller datasets may not form meaningful clusters.
</Warning>

## Dashboard & Visualization Features

The Clustering section provides BuildBetter's full dashboard and visualization capabilities:

* **Real-time Updates**: Dashboards refresh as new signals arrive
* **Interactive Filtering**: Cross-filter across all visualizations
* **Export Options**: Save dashboards as PNG, PDF, or CSV
* **Shareable Links**: Collaborate with teammates on live dashboards
* **15+ Chart Types**: Time series, ridge charts, heatmaps, KPIs, and more
* **Drag-and-drop Builder**: Create custom layouts

<Tip>
  Think of "Clustering" as your signal analytics workspace - it's where you'll find both AI theme grouping AND dashboard/visualization tools.
</Tip>

<Info>
  Clustering transforms overwhelming amounts of feedback into clear, actionable themes, helping you understand what matters most to your customers at scale.
</Info>

The power of clustering lies in its ability to surface insights you didn't know to look for, turning raw feedback into strategic intelligence that drives better decisions.