Analytics

The Analytics page provides interactive visualizations of your feedback data. Use it to spot trends, compare categories, and ask natural-language questions about your insights.

Dashboard

The main dashboard tab shows at-a-glance metrics:

• Volume trends — line chart showing feedback volume over time with a configurable date range
• Sentiment distribution — breakdown of positive, negative, and neutral feedback
• Top categories — bar chart of the most common insight categories
• Value delivery metrics — activation milestones and weekly action throughput
• Top Problems This Week — decision packets for PM + product engineering planning

Drilldown

Click any chart segment to drill down into the underlying insights. For example, clicking the "Bug Report" bar opens a filtered view of all bug reports for that time period.

Decision packets

The Top Problems This Week panel summarizes the top clusters with:

• Weekly evidence count
• Source distribution
• Sync coverage (already synced vs pending)
• A recommended next action draft

Use this section as the agenda input for weekly PM/engineering prioritization.

Explorer

The Explorer tab lets you slice your insights by multiple dimensions:

• By source — group by Slack, Intercom, email, etc.
• By department — group by Sales, Support, Product, Engineering, or Leadership
• By topic — group by AI-identified themes

Use the Explorer to compare feedback patterns across different teams or channels.

AI Query

The AI Query feature lets you ask natural-language questions about your feedback data. Click the AI button on the dashboard to open the query interface.

Example queries

• "How many bug reports came in last week?"
• "What are the top feature requests from Slack?"
• "Show me sentiment trends for the past 30 days"
• "Compare feature requests vs bug reports by source"

How it works

1. You type a natural-language question
2. AI interprets your question and converts it into a structured query
3. The query runs against your feedback database
4. Results are presented as data along with a narrative summary

Query types

The AI can generate four types of queries:

Triage quality and trust

Analytics also includes a Classification Accuracy card with:

• Override rate trend over time
• Confidence-band breakdown (low/medium/high confidence)
• Top override reasons from review decisions
• Prompt/rule tuning recommendations based on recent overrides

These metrics help you improve AI triage quality with a measurable feedback loop.

Proactive alerts

The dashboard shows alert cards for high-signal events such as:

• Negative sentiment spikes
• New critical clusters
• Growing large clusters
• Source failures