> ## 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 Usage Analyst

> Build an AI agent that analyzes product usage, feature adoption, user engagement, and identifies trends

## Overview

A Product Usage Analyst powered by Pylar analyzes product usage patterns, feature adoption rates, user engagement metrics, and identifies trends to inform product decisions.

## What the Agent Needs to Accomplish

The agent must:

* Analyze product usage patterns
* Track feature adoption rates
* Measure user engagement
* Identify usage trends
* Compare user segments
* Recommend feature improvements

## How Pylar Helps

Pylar enables the agent by:

* **Unified Usage View**: Combining product analytics, user data, and feature usage
* **Real-time Analysis**: Querying current usage patterns
* **Trend Identification**: Detecting usage trends over time
* **Segment Comparison**: Comparing usage across user segments
* **Data-Driven Insights**: Generating actionable product recommendations

```mermaid theme={null}
graph LR
    subgraph "Data Sources"
        ANALYTICS[Product Analytics]
        USERS[User Data]
        FEATURES[Feature Usage]
    end
    
    PYLAR[Pylar<br/>Unified Views]
    
    TOOLS[MCP Tools<br/>analyze_usage<br/>track_adoption<br/>identify_trends]
    
    AGENT[Usage Analyst]
    TEAM[Product Team]
    
    ANALYTICS --> PYLAR
    USERS --> PYLAR
    FEATURES --> PYLAR
    PYLAR --> TOOLS
    TOOLS --> AGENT
    AGENT --> TEAM
    
    style PYLAR fill:#FF4017,stroke:#CC3300,color:#fff
```

## Without Pylar vs With Pylar

### Without Pylar

**Challenges**:

* ❌ Multiple analytics tools and data sources
* ❌ Manual usage analysis
* ❌ Difficult to correlate features with engagement
* ❌ Time-consuming trend identification

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

### With Pylar

**Benefits**:

* ✅ Single endpoint for all usage data
* ✅ Automated usage analysis
* ✅ Real-time trend detection
* ✅ Easy feature comparison

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

## Step-by-Step Implementation

### Step 1: Connect Data Sources

1. **Connect Product Analytics** (User actions, events, feature usage)
2. **Connect User Data** (User segments, subscription tiers)
3. **Connect Product Catalog** (Features, releases)

### Step 2: Create Usage Views

**Feature Adoption View**:

```sql theme={null}
CREATE VIEW feature_adoption AS
SELECT 
  f.feature_name,
  f.release_date,
  COUNT(DISTINCT u.user_id) as total_users,
  COUNT(DISTINCT CASE WHEN u.feature_used THEN u.user_id END) as adopted_users,
  (COUNT(DISTINCT CASE WHEN u.feature_used THEN u.user_id END) / 
   NULLIF(COUNT(DISTINCT u.user_id), 0)) * 100 as adoption_rate,
  AVG(u.usage_count) as avg_usage_per_user,
  AVG(u.last_used_days_ago) as avg_days_since_last_use
FROM product.features f
LEFT JOIN analytics.feature_usage u ON f.feature_id = u.feature_id
GROUP BY f.feature_name, f.release_date;
```

**User Engagement View**:

```sql theme={null}
CREATE VIEW user_engagement AS
SELECT 
  u.user_id,
  u.user_segment,
  u.subscription_tier,
  u.active_days_last_30,
  u.features_used_count,
  u.total_actions,
  u.last_activity_date,
  -- Engagement score
  (u.active_days_last_30 * 0.4 + 
   (u.features_used_count / 10) * 30 + 
   (u.total_actions / 100) * 30) as engagement_score
FROM analytics.user_usage u;
```

### Step 3: Create MCP Tools

**Tool 1: Analyze Feature Adoption**

* `analyze_feature_adoption(feature_name: string, days_back: number)`

**Tool 2: Get Usage Trends**

* `get_usage_trends(metric: string, period: string)`

**Tool 3: Compare User Segments**

* `compare_user_segments(segment1: string, segment2: string, metric: string)`

**Tool 4: Identify Usage Patterns**

* `identify_usage_patterns(feature_id: string, user_segment: string)`

## Example Agent Interactions

**User**: "How is the new dashboard feature being adopted?"

**Agent**: "Dashboard Feature Adoption:

* Adoption Rate: 42% (8,400 of 20,000 users)
* Average Usage: 12 times per user
* Top User Segment: Enterprise (65% adoption)
* Growth Trend: +15% week-over-week
* Recommendation: Promote to mid-market segment"

## Outcomes

* **Feature Adoption**: 30% improvement in adoption rates
* **User Engagement**: 25% increase in engagement
* **Product Decisions**: Data-driven feature prioritization
* **Trend Detection**: 60% faster identification of usage trends

## Next Steps

* [Product Feedback Analyzer Example](/examples/product-feedback-analyzer)
* [Feature Flag Manager Example](/examples/feature-flag-manager)