What is MCP natural-language analytics?

**MCP natural-language analytics** lets users ask business questions in plain language. The system connects to MySQL and other databases to return orders, inventory, KPIs and more — without writing SQL or waiting on BI reports.

How it works

1. **Connect data sources**: Bind MySQL in the ScalerDeep console and configure whitelisted tables (e.g. orders, products)

2. **Semantic datasets**: Define how users might ask questions and map them to query logic

3. **Intent routing**: When a user asks, AI matches the best dataset and MCP tool

4. **Results**: Tables, charts, or Excel export for sales reviews and executive decisions

Example questions

• "Sales summary for the last 3 days"

• "Compare order volume by channel this month"

• "Which SKUs have inventory below 10?"

See [MCP query examples](/help/mcp-examples) for more.

MCP vs traditional BI

| Dimension | Traditional BI | MCP natural-language analytics |

|-----------|----------------|--------------------------------|

| Usage | Drag-and-drop dashboards, SQL | Plain-language questions |

| Speed | Analyst queue | Seconds |

| Embedding | Standalone BI | Can embed in AI support and internal assistants |

See [MCP vs traditional BI](/solutions/vs-traditional-bi) for details.

Security

MCP only queries **whitelisted tables** configured by admins. Tenants are fully isolated for enterprise-grade security.