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

# Product Feedback Analyzer

> Build an AI agent that analyzes user feedback, feature requests, sentiment, and prioritizes improvements

## Overview

A Product Feedback Analyzer powered by Pylar analyzes user feedback, feature requests, sentiment, and prioritizes product improvements based on data-driven insights.

## What the Agent Needs to Accomplish

The agent must:

* Analyze user feedback and reviews
* Track feature requests
* Perform sentiment analysis
* Prioritize improvements
* Identify common pain points
* Track feedback trends

## How Pylar Helps

Pylar enables the agent by:

* **Unified Feedback View**: Combining feedback from multiple sources
* **Sentiment Analysis**: Automated sentiment scoring
* **Priority Calculation**: Data-driven prioritization
* **Trend Analysis**: Identifying feedback patterns over time

```mermaid theme={null}
graph LR
    subgraph "Data Sources"
        FEEDBACK[User Feedback]
        REVIEWS[Reviews]
        REQUESTS[Feature Requests]
    end
    
    PYLAR[Pylar<br/>Unified Views]
    
    TOOLS[MCP Tools<br/>analyze_sentiment<br/>prioritize_features<br/>track_trends]
    
    AGENT[Feedback Analyzer]
    TEAM[Product Team]
    
    FEEDBACK --> PYLAR
    REVIEWS --> PYLAR
    REQUESTS --> PYLAR
    PYLAR --> TOOLS
    TOOLS --> AGENT
    AGENT --> TEAM
    
    style PYLAR fill:#FF4017,stroke:#CC3300,color:#fff
```

## Without Pylar vs With Pylar

### Without Pylar

**Challenges**:

* ❌ Feedback scattered across platforms
* ❌ Manual sentiment analysis
* ❌ Difficult to prioritize requests
* ❌ Limited trend visibility

**Implementation Complexity**: \~4-5 weeks

### With Pylar

**Benefits**:

* ✅ Single endpoint for all feedback
* ✅ Automated sentiment analysis
* ✅ Data-driven prioritization
* ✅ Real-time trend tracking

**Implementation Complexity**: \~5-6 hours

## Step-by-Step Implementation

### Step 1: Connect Data Sources

1. **Connect Feedback Platform** (UserVoice, Canny, Intercom)
2. **Connect Reviews** (App Store, G2, Capterra)
3. **Connect Support** (Support tickets, feature requests)

### Step 2: Create Feedback Views

**Feedback Analysis View**:

```sql theme={null}
CREATE VIEW feedback_analysis AS
SELECT 
  f.feedback_id,
  f.feedback_text,
  f.category,
  f.votes,
  f.sentiment_score,
  f.created_date,
  -- Priority score
  (f.votes * 0.4 + 
   ABS(f.sentiment_score) * 30 + 
   CASE WHEN f.category = 'Critical' THEN 30 ELSE 0 END) as priority_score
FROM feedback.feedback f;
```

### Step 3: Create MCP Tools

**Tool 1: Analyze Feedback Sentiment**

* `analyze_feedback_sentiment(category: string, days_back: number)`

**Tool 2: Get Top Feature Requests**

* `get_top_feature_requests(limit: number, min_votes: number)`

**Tool 3: Identify Common Pain Points**

* `identify_pain_points(category: string, min_mentions: number)`

**Tool 4: Prioritize Improvements**

* `prioritize_improvements(criteria: string, limit: number)`

## Example Agent Interactions

**User**: "What are users saying about our mobile app?"

**Agent**: "Mobile App Feedback Analysis:

* Overall Sentiment: 3.8/5 (Moderately Positive)
* Top Issues: App crashes (45 mentions), Slow loading (32 mentions)
* Top Requests: Dark mode (120 votes), Offline mode (95 votes)
* Priority: Fix app crashes (High), Add dark mode (Medium)"

## Outcomes

* **Feedback Processing**: 70% faster analysis
* **Prioritization**: 40% improvement in feature prioritization
* **User Satisfaction**: 25% increase in satisfaction scores
* **Product Roadmap**: Data-driven roadmap decisions

## Next Steps

* [Product Usage Analyst Example](/examples/product-usage-analyst)
* [Feature Flag Manager Example](/examples/feature-flag-manager)