An analysis and a pilot project, conducted on behalf of the Ministry of Economic Affairs and Communications, were completed to explore the automation of directing enquiries from citizens and businesses to the appropriate authorities and officials. Thanks to the project, people will not have to know which specific authority or official to contact in the future. Siimuke thus helps to reduce the time spent to respond to enquiries and serves as a precondition for the adoption of Bürokratt across public authorities.

The kratt helps doctors write down examination results through dictation. The vocabulary used in examination descriptions is generally very specific and fairly predictable. The whole essence of speech recognition technology is based on probability. The software listens to the person's speech and tries to find which written word is most likely to match the sound it hears. Medical jargon may be difficult for the average person, but it has a more limited range of words, which makes it more likely that the software will spell it correctly.

Sensors and AI derive the salinity of the road surface and help save resources.

The National Heritage Board developed a new prototype called Sälli that helps preserve museum artifacts in the best way. A forecast model that helps predict the ageing of museum artifacts was created.

Texta Toolkit is an open-source toolkit for accessing and analysing free text (big) data. The software is designed to process information from different areas. The Ministry of Education and Research performed a document management audit using the Texta Toolkit, which was aimed at identifying the documents that had been published without authorisation (eg internal documents, personal data etc).

The processing of emergency notifications is an extensive and complex tasks, which can lead to dispatchers making mistakes and/or spending too much time providing their threat assessment. When providing a threat assessment, the need-to-know of at least three different agencies must be taken into consideration to dispatch resources, while also processing the emergency call within one minute. Gathering information in a time-constrained environment is challenging for people, thus reducing errors is of critical importance. The threat assessment tool of the Emergency Response Centre is an AI-based software application that helps to provide the dispatcher with the most likely scenario by transcribing the incoming call and taking into account previous actions.

The tool detects potential VAT fraud. The purpose of the project was to identify inaccurately filed VAT refund claims, whether done intentionally (fraud) or due to human error. The developed machine learning model enables scoring of future VAT refund claims, allowing the determination of the probability of whether a refund claim is correctly or incorrectly filed.

The Estonian Tax and Customs Board leverages AI for more efficient tax collection. The tool for detecting possible envelope wage payers identifies potential payers of envelope wages using a risk model.

A kratt for analysing conversations with customers at The Estonian Health Insurance Fund.

The project used machine vision to identify road signs.

A kratt for transcribing and analysing customer chats in the Unemployment Insurance Fund.

As part of the project, crisis hotline calls were transcribed and analysed to collect data for the development of future kratts.