Digital assistant of the Ministry of Labour and Social Affairs – chatbot and voicebot for public communication

Citizens often struggle to navigate the complex agendas of the Ministry of Labour and Social Affairs, while call centers are overloaded with a high volume of inquiries. Manual handling of questions is time-consuming, increases administrative burden, and reduces efficiency. The project addresses this challenge through an AI assistant that automates responses to frequently asked questions and enables quick escalation of more complex cases to human operators.

The project introduces a digital assistant for the Ministry of Labour and Social Affairs in the form of a chatbot and voicebot to provide citizens with verified public information. The assistant leverages retrieval-augmented generation (RAG), semantic search, and speech recognition for fast, consistent, and accessible communication. The goal is to provide citizens with reliable 24/7 answers, reduce the workload of call centers, and improve the efficiency of government–citizen communication.

15

Python, LLM, Vector DB, Azure

Business Architect

Complete realization

Availability according to the office hours set by the Labour Office. Faster citizen access to verified information.

Reduction of call center administrative workload by up to 40%.

Faster citizen access to verified information.

The system is already in production, achieving high accuracy with an error rate below 5%. The digital assistant has processed tens of thousands of citizen inquiries. The deployment has significantly reduced the burden on call centers and has been positively received by both the public and the MoLSA staff.

Implementation of the voicebot and technical integration of all components to achieve optimal fluency. Consolidation of unstructured documents and creation of a knowledge base. Implementation of the RAG system.

In our practice of deploying AI systems to support organizations, we often encounter situations where clients attempt to address or optimize their challenges and processes using commonly available tools. However, these tools are usually designed as universal solutions, which in practice often leads to issues with accuracy, reliability, and overall efficiency. Our goal is to prevent such shortcomings already at the solution design stage. Thanks to our experience and proven methodologies, we are able to guide the client smoothly and effectively through the entire implementation process—from needs analysis to the deployment of tailored tools.