How we replaced 10 analysts with a single tool
While building an ERP system we discovered a need to streamline the analytics team's workflow — and solved it with AI.
The project
While developing an ERP system for a company specializing in metal-cutting tooling and industrial equipment, we came up with an idea to try optimizing the work of the company's analytics department.
Pegas has been successfully supplying metal-cutting tooling and equipment since 2006. Over time, the company has made a major leap forward and now offers a full set of services — from selling tooling and equipment to introducing your parts onto the production line. Beyond tool sales, the company also runs an engineering-services line: starting from the customer's drawing all the way to producing finished products on your own equipment.
Solution
We built a query handler and connected the company's internal databases — Orders, Customers, Sales, Production, Logistics, Reports — to a large language model (LLM). This let users query the system directly with analytical questions in plain conversational language. For example, to check current sales-plan performance for the quarter the user simply asks: "What's the current sales plan execution status — give me the answer in rubles and in dollars at today's exchange rate." The system pulls data from the sales database, fetches today's USD rate from the web, runs the math, and replies in a clear format within seconds.
Result
The tool remains the most popular module among employees. The analytics team no longer spends time on small ad-hoc requests from colleagues and is now fully focused on large initiatives.
