This guide shows you how to enable emotional analysis in Togetherly to track changes in service user wellbeing over time based on AI analysis of call recordings.

## Prerequisites

- Your organization must have call recordings enabled in settings
- Your account must have the Conversational Analysis feature enabled
- At least one recorded call must be completed for analysis results to appear

## Steps

### Step 1: Enable emotional analysis in account settings

Navigate to Settings from the main menu, then select Account Settings. In the 'Enable conversational analysis' section, first ensure the main conversational analysis toggle is switched on. Then scroll down and check the box next to 'Emotional analysis' which states 'Measure changes in service user emotion over time.' Click 'Save Settings' at the bottom of the page to apply your changes. The system will now automatically analyze the emotional tone of all future recorded calls.

![Enable emotional analysis in account settings](https://help.together.ly/rails/active_storage/blobs/proxy/eyJfcmFpbHMiOnsiZGF0YSI6MTEwLCJwdXIiOiJibG9iX2lkIn19--ed5b79686a3f4066075b024cf5077c1f211879a9/step_0.png)

### Step 2: Review mood indicators for individual calls

Once emotional analysis is enabled, mood indicators will appear in your befriending match call history tables. Navigate to the Befriending section and select any active match to view its call history. In the table, you'll see a 'Mood' column displaying emoji icons that represent the AI-detected emotional state during each call: a smiling face for happy, a frowning face for sad, an angry face for stressed or angry, and a neutral face for neutral emotions. These indicators help you quickly identify patterns in a service user's emotional wellbeing across multiple interactions.

![Review mood indicators for individual calls](https://help.together.ly/rails/active_storage/blobs/proxy/eyJfcmFpbHMiOnsiZGF0YSI6MTExLCJwdXIiOiJibG9iX2lkIn19--29d379011101a03ae8b17b738960f2aba991f5ca/step_1.png)

### Step 3: Access detailed wellbeing reports

To view wellbeing trends over time, navigate to the Reports section from the main menu and select 'Impact Report.' Find the service user you want to review and click on their name to access their individual report. The wellbeing report displays a bar chart showing the service user's emotional wellbeing score tracked month by month, along with a current wellbeing score and percentage change indicator. This visualization helps you identify whether a service user's emotional state is improving, declining, or remaining stable, enabling you to make informed decisions about their care and support needs.

![Access detailed wellbeing reports](https://help.together.ly/rails/active_storage/blobs/proxy/eyJfcmFpbHMiOnsiZGF0YSI6MTEyLCJwdXIiOiJibG9iX2lkIn19--f59f1dd97329ee7c788d1c18d45245d70f36b850/step_2.png)

### Step 4: Interpret emotional analysis results

The AI analyzes call transcriptions to assess service user emotions across five categories: happy, sad, angry, stressed, and neutral. The wellbeing score is calculated based on the frequency and intensity of positive versus negative emotional indicators throughout conversations. A score above 70 generally indicates good emotional wellbeing, while scores below 50 may suggest the service user needs additional support. Remember that emotional analysis provides helpful insights but should always be considered alongside other information and professional judgment when making decisions about care.

## Tips

- Emotional analysis works best with calls lasting at least 3 minutes, as shorter calls may not provide enough conversational context for accurate assessment
- Review wellbeing trends regularly to identify gradual changes that might not be immediately obvious from individual call interactions
- Combine emotional analysis data with safeguarding alerts and call summaries for a comprehensive understanding of each service user's wellbeing

## Summary

By enabling emotional analysis, you can track service user wellbeing trends over time through AI-powered mood detection in call recordings. This helps you identify when service users may need additional support and measure the positive impact of your befriending services.