Skip to content
AI Lab
Contact

AI Analytics

Ask a question in plain English and get the query, a chart, and a written summary.

Simulated
Natural language to insight
  • Text-to-SQL
  • Charting
  • Summaries
Acme sales warehouse (sample)Ready

Ask a question about Acme's sales data

Write it the way you'd ask an analyst. The schema panel lists every table and column you can ask about.

  1. 1Translate to SQLThe question is mapped to tables and columns, and written as a query you can read.
  2. 2Run on the warehouseRead-only query against the sample data, with row counts and timing.
  3. 3Chart and summarizeThe best chart for the shape of the result, the rows, and a short written summary.

How it works in production

The demo above runs on a script in your browser. This is the architecture it stands in for.

Question

Natural language

Semantic layer

Shared metric definitions

Generation

Text-to-SQL

Schema-aware prompt

SQL validator

Read-only, row limits

Execution

Warehouse

BigQuery or Postgres

Presentation

Chart selection

Written summary

Saved reports

Question to insight

The model writes SQL against a semantic layer of agreed metric definitions, not raw tables, so "revenue" means the same thing in every answer. Generated queries are parsed and checked (read-only, allowed tables, row limits) before they run.

The query, chart and summary are shown together so anyone can check how a number was produced. Useful questions can be saved as reports and scheduled.

Build something like this

Tell me about the process you want to improve and the systems it touches. I'll come back with questions and a suggested approach.

Build something like this