Skip to main content
Brussels Capital-Region - Homepage

"Chatbot and AI in public service" experiment: Mission accomplished!

In 2024, Paradigm launched an experiment to explore the technological advances of large language models (LLM) for the benefit of public service.

Published on

The aim of this exploratory study was to determine specifically whether, in the case of the External linkIRISbox electronic service counter, a chatbot-type conversational assistant using generative AI could serve as an interactive guide to navigate through the various online citizen services that the Region offers.

The aim was to evaluate whether a chatbot using generative AI is capable of streamlining online procedures, thus freeing up the relevant administrations from certain repetitive tasks. 

This experiment was carried out in a "sandbox" environment (which enables features to be tested without affecting the production environment) and is the result of collaboration between Paradigm's IRISbox team, IBM and NRB.

Lessons learnt

The experiment proved conclusive:

  • The use of a knowledge base had an impact on the quality of the accurate and relevant information provided by the chatbot.
  • The infrastructure chosen (IBM Watson) guarantees data confidentiality.
  • The chatbot detects the language the person is using and responds in the same language.
  • The chatbot communicates information specific to each municipality, when the name of the municipality has been provided; there is an integrated link to the IRISbox procedure and, once connected, a contextual mini guide may appear. 

Benefits of an exploratory project

This experiment involved one specific administrative procedure out of the hundred or so processes that IRISbox offers. 

In 2024, Paradigm also carried out an exploratory project for a support chatbot for civil servants, at the request of parking.brussels, supported by Salesforce and Deloitte.

The advantage of such exploratory projects is that they enable us to assess the potential of new technologies on a limited scale, and to develop the skills of the teams involved in concrete cases. Opting for an exploratory project can therefore help you to map out a future project by identifying challenges, reducing risks and cutting unnecessary costs.

Back to top