Reports

Reports are where you chat with your data. Ask questions in natural language and get instant analytics powered by MetricChat's intelligent Agent system.

How It Works

Type a question in the prompt box and MetricChat's agent analyzes your data to provide answers with visualizations, tables, and explanations.

Prompt Types

• Query-style — Request specific metrics: "Show me total revenue by month"
• Analysis-style — Investigate trends: "Why did sales drop in March?"
• Comparative — Compare segments: "How does Q4 performance compare to Q3?"

Result Formats

The agent provides:

• Data tables with raw results
• Auto-generated visualizations (charts, graphs)
• Executable SQL or Python code
• Written explanations and insights

What You Can Do With Results

• Export as CSV or Excel
• Modify the generated code
• Add visualizations to dashboards
• Share with team members

Conversational Iteration

Questions build on previous responses. Start with a broad question and refine:

1. "Show me monthly revenue" → line chart
2. "Break it down by product category" → stacked chart
3. "Just the top 5 categories" → filtered view
4. "Add year-over-year comparison" → comparative analysis

Agent Reasoning

Behind every answer, the agent follows a structured loop:

1. Think — Plan the approach
2. Act — Search context, generate queries, build visualizations
3. Observe — Validate results, check for errors
4. Reflect — Determine if the answer is complete or needs refinement

Tips for Better Results

• Be specific — "Active users in Q4 2024" beats "show me users"
• Include business context — Mention relevant terms, metrics, or filters
• Leverage instructions — Define KPIs and rules so the AI applies them automatically
• Review generated code — Transparency into exactly how results were produced
• Choose the right LLM — Select from the dropdown based on task complexity