Case study
Querying a database in natural language.
Clait built a conversational business intelligence platform: staff type a question in plain language, the system turns it into an optimised SQL query and returns an interactive chart, with no SQL knowledge needed.
Published
01
The context
Managing a complex database of courses, participants and performance metrics often takes technical skills and a long wait for every single analysis. The goal: make the data accessible to everyone, in real time.
02
The challenges
Reliance on IT
Every chart or analysis meant writing an SQL query by hand.
Long turnaround
Analyses were not instant, which slowed decisions down.
Technical barrier
Only people who knew SQL could get insights from the data.
Rigidity
Every new question needed ad hoc development.
Operational scale
More requests meant more load on IT.
03
The solution
A conversational business intelligence platform that turns the database into an intelligent tool, in four steps.
Plain-language chat
Staff type their questions in natural language.
Translation into SQL
The AI engine picks the tables and columns, works out aggregation, filtering and sorting, and generates correct, efficient SQL.
Chart-ready data
The data is reshaped for visualisation; the chart type, from multi-line to scatter to heatmap, is chosen from the shape of the data and the request, within the interface’s limits.
Interactive charts
The result is a dynamically generated, interactive chart.
04
Technology
- Frontend
- Vue.js 3, D3.js, Chart.js, amCharts
- Backend
- Django, FastAPI
- LLM
- OpenAI models through Azure AI
- Database
- PostgreSQL
- Infrastructure
- Azure VPC
- Languages
- Python (LLM), TypeScript (frontend)
FAQ
Frequently asked questions.
Do users need to know SQL?
No. The user types the question in plain language; the system generates the SQL query and returns the chart.
How does it choose the chart type?
From the shape of the extracted data and the user’s request, within the interface’s dimensionality limits: multi-line, scatter, heatmap and more.
What infrastructure does it run on?
An Azure VPC, with OpenAI models provided through Azure AI, a Django and FastAPI backend and a PostgreSQL database.
Keep reading.
- Service · 01AI agentic systems for complex processesClait develops LLM-based agentic systems: specialised AI agents that work together to orchestrate complex workflows, make decisions, correct themselves and work with the data and tools a company already uses.Agentic systems · Works with your tools · Automation
- Case study · 02AI agents for company researchBrandy is an internal Clait project: starting from a company’s website, a multi-agent system gathers information from public sources, structures it into a report and drafts strategies and first-contact messages, supporting lead generation and prospecting.Agentic systems · Lead generation
Contact
We find a solution for everyone.
We'll find yours too.
Tell us what you'd like to improve, even if for now it's only an idea. We'll start from there and build the rest together.