LLM Agents for Automated Extraction of Supplier Pricing from Email Workflows

Design and prototyping of an agentic system for acquiring product costs from suppliers using LLM-based agents

Requirements

  • M.Sc. in Machine Learning, Data Science, Computer Science, Mathematics, Telecommunications, or similar
  • Knowledge of Python
  • Software development skills
  • Basic concepts of LLMs and agents
  • Basic knowledge of frameworks like LangChain

Description

A local company intends to tackle the challenge of automating the process of acquiring and updating the costs of products supplied by external partners. Currently, this process is handled mainly through email exchanges between the company and its suppliers, followed by manual or semi-manual updates of the internal database. This approach has shortcomings in terms of time, scalability, risk of error, traceability of information, and standardization of operational workflows. In this scenario, the recent evolution of large language models (LLMs) and agentic systems offers new opportunities to automate significant parts of the process, improving the ability to interpret received communications, extract relevant data, validate information, and support integration with the company’s information systems.

The internship aims to study, select, configure, and prototype an intelligent solution based on LLM agents to support the process of acquiring product costs from suppliers, from email communications up to the structured update of the internal database. The work will focus both on analyzing the current process and on defining a prototype architecture capable of:

  • reading and interpreting emails received from suppliers;
  • identifying and structuring the relevant cost information;
  • handling any ambiguities or missing data through control logic or interaction;
  • preparing the update of the company database in a consistent and traceable manner;
  • supporting company operators through supervised workflows.

By the end of the internship, the expected outcomes are:

  • a clear mapping of the current process and of the opportunities for improvement;
  • a feasibility study on the use of LLM agents for the company’s use case;
  • an architectural proposal for the intelligent automation of the process;
  • an initial prototype in a simplified setting, capable of demonstrating the automated or semi-automated handling of email requests/receipts and the structuring of cost data.

Educational value of the internship

The internship will offer the opportunity to work on a real case of business innovation, combining process analysis, generative artificial intelligence, agentic system design, and software prototyping. The student will be able to acquire practical skills in applying LLM technologies to concrete problems of document automation and decision-making support.

Contacts