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

# Research

> Browse and analyze customer conversations from Granola meetings and manual transcripts

Research is a powerful feature within Products that helps you understand customer needs by browsing meeting transcripts, exploring insights, and connecting customer signals to product decisions.

## Overview

The Research tab provides a conversation navigator that displays:

* **Meeting transcripts** from connected Granola accounts
* **Manual transcripts** pasted directly into Kasava
* **AI-powered exploration** to find patterns across conversations

Navigate to `/products/[id]/analysis?analysisTab=insights` and select the Research subtab to access this functionality.

## Conversation Navigator

The Research interface uses a two-panel layout:

* **Left panel**: Searchable list of meetings with metadata
* **Right panel**: Full meeting details with transcript and actions

### Meeting List

Each meeting in the list displays:

* **Title** - Meeting name or calendar event title
* **Summary** - AI-generated meeting summary (if available)
* **Timestamp** - When the meeting occurred
* **Duration** - Meeting length in minutes
* **External attendees** - Badge showing external participant count
* **Action items** - Count of extracted action items
* **Source badge** - Indicates if the transcript is from Granola or manually added

### Filtering and Search

Use the search bar to filter meetings by:

* Meeting title
* Summary content
* Participant names
* Discussed topics

## Adding Manual Transcripts

Not all users have access to Granola's enterprise features. Kasava supports adding transcripts manually from any source.

### Opening the Dialog

1. Click the **Add Transcript** button in the Research header
2. Or click **Add Transcript** in the empty state when no meetings exist

### Adding a Transcript

<Steps>
  <Step title="Enter Title">
    Provide a descriptive title for the meeting or conversation
  </Step>

  <Step title="Paste Transcript">
    Paste the full transcript text into the textarea. The transcript should be at least 10 characters.
  </Step>

  <Step title="Optional Details">
    Click "Optional details" to expand additional fields:

    * **Source** - Where the transcript came from (Manual, Granola, Otter.ai, Zoom, Other)
    * **Date** - When the meeting occurred
    * **Duration** - Meeting length in minutes
  </Step>

  <Step title="Save">
    Click **Add** to save the transcript. It will appear immediately in the meeting list.
  </Step>
</Steps>

### Supported Sources

| Source   | Description                               |
| -------- | ----------------------------------------- |
| Manual   | Typed or pasted from any source           |
| Granola  | Copied from Granola meeting notes         |
| Otter.ai | Exported from Otter transcription service |
| Zoom     | Exported from Zoom meeting transcripts    |
| Other    | Any other transcription tool              |

<Tip>
  When pasting from Granola, select "Granola" as the source to help with future analytics and source tracking.
</Tip>

## Meeting Detail View

Click any meeting to view its full details in the right panel.

### Detail Sections

* **Header** - Meeting title, date, duration, and source indicator
* **Summary** - AI-generated meeting summary
* **Transcript** - Full conversation transcript with speaker labels
* **Action Items** - AI-extracted tasks, decisions, and follow-ups with export capabilities
* **Attendees** - List of participants with internal/external indicators

### Transcript Features

The transcript view supports:

* **Text selection** - Select text to explore related insights
* **Copy** - Copy transcript content for use elsewhere
* **Search** - Find specific phrases within the transcript

## AI Exploration

Select text within a transcript to trigger AI-powered exploration that finds:

* **Similar signals** - Other conversations mentioning related topics
* **Related metrics** - Data points connected to the discussed topic
* **Suggested metrics** - New metrics you could track based on the conversation
* **Feature ideas** - Product features that could address discussed needs

### Starting an Exploration

1. Select text within a transcript
2. Click the **Explore** button that appears
3. Wait for AI analysis to complete
4. Review results in the exploration panel

<Note>
  Explorations run asynchronously. You can continue browsing while the AI analyzes your selection.
</Note>

### Exploration Results

| Result Type       | Description                                                   |
| ----------------- | ------------------------------------------------------------- |
| Similar Signals   | Past conversations with related content, ranked by similarity |
| Related Metrics   | Existing metrics connected to the topic                       |
| Suggested Metrics | New metrics the AI recommends creating                        |
| Feature Ideas     | Product feature suggestions with complexity estimates         |

## AI Action Item Extraction

Kasava automatically analyzes meeting transcripts to extract actionable items, decisions, questions, blockers, and follow-ups using advanced AI.

### How It Works

When you upload a transcript or sync a meeting from Granola, Kasava uses Claude AI to:

1. **Identify action items** - Tasks that need to be completed
2. **Capture decisions** - Decisions made during the conversation
3. **Extract questions** - Unanswered questions that need follow-up
4. **Detect blockers** - Issues preventing progress
5. **Find follow-ups** - Items requiring future discussion

### Analyzing a Transcript

<Steps>
  <Step title="Select a Meeting">
    Click any meeting in the Research list to view its details
  </Step>

  <Step title="Click Analyze">
    In the meeting detail panel, click the **Analyze Transcript** button
  </Step>

  <Step title="Review Results">
    AI extraction typically completes in 10-30 seconds. Results appear in the Action Items section.
  </Step>
</Steps>

### Action Item Details

Each extracted action item includes:

| Field                  | Description                                            |
| ---------------------- | ------------------------------------------------------ |
| **Title**              | Concise, actionable description (max 100 characters)   |
| **Description**        | Detailed context from the transcript                   |
| **Priority**           | Critical, High, Medium, or Low (inferred from urgency) |
| **Category**           | Task, Decision, Question, Blocker, Follow-up, or Other |
| **Mentioned Assignee** | Person's name if mentioned in the conversation         |
| **Confidence Score**   | AI confidence level (0-100) in the extraction accuracy |
| **Extracted From**     | Relevant quote from the transcript                     |
| **Keywords**           | 3-5 relevant keywords for searchability                |

<Tip>
  **Priority Guidelines**: The AI infers priority based on language cues:

  * **Critical** - Blockers, urgent deadlines, security issues
  * **High** - Important work with near-term deadlines
  * **Medium** - Standard tasks, most decisions
  * **Low** - Nice-to-haves, low-urgency items
</Tip>

## Exporting Action Items

Convert AI-extracted action items into tickets in your task management platform with a single click.

### Supported Platforms

Export action items to any of these platforms:

| Platform   | Export Type   | Requirements                                                  |
| ---------- | ------------- | ------------------------------------------------------------- |
| **Linear** | Issue         | Connected Linear integration                                  |
| **Jira**   | Issue         | Connected Jira integration                                    |
| **Asana**  | Task          | Connected Asana integration                                   |
| **GitHub** | Issue         | Repository access                                             |
| **Notion** | Database Page | Connected Notion workspace with configured ticketing database |

### Exporting to Linear, Jira, or Asana

<Steps>
  <Step title="Select Action Item">
    In the Action Items section, click the action item you want to export
  </Step>

  <Step title="Click Export">
    Click the **Export** button and select your platform
  </Step>

  <Step title="Configure Details">
    * **Linear**: Select team, project (optional), and assignee
    * **Jira**: Select project, issue type, and epic (optional)
    * **Asana**: Select project, section, and assignee
  </Step>

  <Step title="Create Ticket">
    Click **Export** to create the ticket. A link to the external ticket will be saved with the action item.
  </Step>
</Steps>

### Exporting to Notion

Notion's flexible database structure requires one-time configuration before you can export action items.

#### Setting Up Notion Ticketing

<Steps>
  <Step title="Connect Notion">
    First, ensure your Notion workspace is connected in **Settings** > **Integrations** > **Notion**
  </Step>

  <Step title="Create a Ticketing Database">
    In Notion, create a database for action items with the properties you want to track (Title, Priority, Status, etc.)
  </Step>

  <Step title="Configure Ticketing in Kasava">
    Navigate to **Settings** > **Integrations** > **Notion** > **Ticketing Configuration**
  </Step>

  <Step title="Select Database">
    Choose the Notion database where action items should be created
  </Step>

  <Step title="Map Properties">
    Map Kasava fields to your Notion database properties:

    * **Title Property** - Which Notion property to use for the action item title (required)
    * **Description Property** - Where to store the detailed description
    * **Status Property** - Map status values (Open → "To Do", In Progress → "Doing", etc.)
    * **Priority Property** - Map priority values (Critical → "Urgent", High → "High", etc.)
    * **Category Property** - Map category values (Task → "Task", Decision → "Decision", etc.)
    * **Source Property** - Property to indicate the item came from a transcript
  </Step>

  <Step title="Save Configuration">
    Click **Save** to complete the setup
  </Step>
</Steps>

<Note>
  **Property Mapping**: Your Notion database can use any property names. Kasava maps Kasava's standard fields to your custom property names, making it flexible for any database schema.
</Note>

#### Exporting Action Items to Notion

Once configured, exporting to Notion works the same as other platforms:

1. Select an action item from the transcript
2. Click **Export** and choose **Notion**
3. The action item is created as a page in your configured database
4. A link to the Notion page is saved with the action item

<Tip>
  **Multiple Configurations**: You can configure different Notion databases for different organizations. Each organization can have its own ticketing setup even when using the same Notion workspace.
</Tip>

### Export Tracking

Once exported, action items display:

* **Platform badge** - Shows which platform the item was exported to
* **External link** - Direct link to the ticket in the external platform
* **Export timestamp** - When the export occurred
* **Export status** - Success or error state

<Warning>
  **Duplicate Prevention**: Each action item can only be exported once per platform. To create multiple tickets, extract the action item again or duplicate it manually in the target platform.
</Warning>

## Connecting to Granola

For automatic meeting sync, connect your Granola account:

<Steps>
  <Step title="Open Empty State">
    Navigate to Research with no meetings linked
  </Step>

  <Step title="Click Connect Granola">
    Click the **Connect Granola** button
  </Step>

  <Step title="Complete OAuth">
    Follow the Granola authorization flow
  </Step>

  <Step title="Configure Sync">
    Set your sync preferences in the integration settings
  </Step>
</Steps>

Alternatively, navigate to **Settings** > **Integrations** > **Granola** to connect.

<Warning>
  **Granola Enterprise Required**: Automatic meeting sync via API is only available to Granola Enterprise customers. Users on other plans can still paste transcripts manually using the Add Transcript feature.
</Warning>

## Best Practices

<Tip>
  **Add context-rich titles** - Use descriptive titles for manual transcripts to make them easier to find later.
</Tip>

<Tip>
  **Include meeting dates** - Setting the meeting date helps with chronological organization and trend analysis.
</Tip>

<Tip>
  **Explore key quotes** - Select impactful customer quotes to find related conversations and identify patterns.
</Tip>

<Tip>
  **Link to decisions** - When a conversation leads to a product decision, log it in the Decisions tab and reference the conversation.
</Tip>

## Related

<CardGroup cols={2}>
  <Card title="Products" icon="box" href="/docs/features/products">
    Overview of product management in Kasava
  </Card>

  <Card title="Granola Integration" icon="calendar" href="/docs/integrations/granola">
    Connect Granola for automatic meeting sync
  </Card>

  <Card title="AI Chat" icon="message-circle" href="/docs/features/ai-chat">
    Query your research data using natural language
  </Card>

  <Card title="Insights" icon="lightbulb" href="/docs/features/products#insights-ai-analysis">
    AI-generated product analysis
  </Card>
</CardGroup>