Natural Language BI Agent

Ask your questions in plain English, get your data in seconds

WhatsApp
Agent sheet
Category
BI and decision intelligence
Model
open, installed on your premises
Data
never leaves
Subscription
€0 per month
01Overview

Traditional business intelligence creates analyst dependency: every new question requires a request, a wait and a partial answer. The Natural Language BI Agent removes this intermediary. Ask your question in English or French, the agent generates the corresponding SQL, executes it on your data warehouse and returns a table or chart in seconds.

The problem

Business teams depend on data teams for every query. Data analysts are overloaded with ad hoc requests. Static dashboards don't answer questions that arise in meetings. And decisions are made without data, or with outdated data.

Our answer

The agent is connected to your data warehouse (BigQuery, PostgreSQL, Snowflake, Redshift). It translates your questions into valid SQL, executes it in real time and returns results as a table, chart or text summary. A question history is maintained to feed a self-learning dashboard.

02How it works

Live in four steps.

  1. 01
    Data warehouse connection

    BigQuery, PostgreSQL, Snowflake, Redshift or MySQL integration. Automatic schema documentation for the agent.

  2. 02
    Business vocabulary calibration

    LLM context enrichment with your specific vocabulary: table names, metrics, KPIs, calculation rules.

  3. 03
    Query interface

    Simple chat interface (Slack bot, web app or widget) where teams ask questions in natural language.

  4. 04
    Self-learning dashboard

    Frequent questions feed a Metabase or Power BI dashboard updated automatically. Continuous learning.

03What you gain

Results you can measure.

Data autonomy for everyone

Any team member can query data without writing a line of SQL or waiting for an analyst.

Fact-based decisions

KPIs are accessible in real time during meetings, with no delay. Decisions are based on real data.

Data analysts freed up

Your analysts stop spending 60% of their time on ad hoc requests. They focus on strategic analysis.

04Frequently asked

What we get asked before signing.

Which data warehouses are supported?
BigQuery, PostgreSQL, MySQL, Snowflake, Redshift, SQLite. Other sources can be added via connectors.
Can the agent handle complex schemas with dozens of tables?
Yes. The agent automatically documents your schema and learns table relationships through example queries.
Can results be exported?
Yes. Each result can be exported as CSV, integrated into Notion or pushed to Slack automatically.
How is data security ensured?
The agent runs in your infrastructure (VPC or on-premise). Data never leaves your environment.

Let's talk about your case for thirty minutes.

A conversation to understand your context, volumes and tools. You leave with a scope and an order of magnitude, no commitment.

WhatsApp
+261 38 66 261 00team@wikolabs.com