Using a machine learning model trained on the data of unemployed persons from the past five years, OTT, the decision support of the Unemployment Insurance Fund, summarises the situation of a specific person by predicting the probability of them finding employment during the year and the probability of them experiencing unemployment again, highlighting the factors affecting these probabilities. In this way, counsellors can get a quick overview of a customer's situation, along with an overview of all their customers based on OTT assessments, which allows them to set priorities according to the customers' need for assistance. In addition, OTT assessments are also used by the managers of counsellors to distribute workloads more evenly and provide support to counsellors with more complex customer portfolios.

A website that can be used to automatically transcribe audio files containing Estonian speech.

An application was created that makes the identification and analysis of the pay gap as simple and accessible as possible for businesses, without increasing the administrative burden. Additionally, the application promotes awareness and transparency among businesses/organisations regarding pay inequalities and the pay gap. The Pay Mirror platform enables an automated analysis of pay gap indicators in organisations based on register data and provides an opportunity to understand compliance with the principles of equal pay in the organisation and to improve the situation if necessary. Furthermore, this digital solution aids employers in elucidating the pay gap through objective, gender-neutral factors and, when necessary, in identifying any potential unjustified discrepancies. Supervision can also be partly automated with the help of this digital solution. The digital solution can be used by all employers regardless of the number of employees.

Personalised learning involves a transition from a teacher-centric to a learner-centric learning process where learners have the opportunity to make choices based on their own goals, pace, tools and interests. The objective is to utilise the available data to develop solutions that provide feedback to both Estonian learners and teachers on the learning process and to even go beyond what has already been done by showcasing what the learner has accomplished, explain their actions and identify potential gaps in their knowledge or skills.

As part of the project, prototypes were developed based on aerial laser scanning and drone surveillance. They are used to map and monitor construction sites. Sample procedures are provided for both buildings and road construction. The aerial laser scanning prototype relies on elevation data from the Land Board and its application is focused on the different stages of a construction project. Drone surveillance prototypes rely on the data collected as part of the project.

Determining the degree and thickness of snow cover using AI.

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. In collaboration with the Centre of Registers and Information Systems, the Ministry of Justice removed, using Texta, personal data from nearly 80,000 court decisions involving deleted punishments, and republished the decisions in the Court Information System.

In co-operation with computer scientists, the neurorehabilitation research group at the Children's Clinic of Tartu University Hospital is developing a robot therapist that could alleviate the difficulties with social skills of Estonian children. When children interact with a robot, they do not experience the social anxiety or shyness that often arises when interacting with adults. This allows children to concentrate more effectively on developing the content of their speech. The introduction of the robot for medical purposes allowed the successful training of children's conflict resolution and self-expression skills, emotions, vocabulary and manners.

The ARIB program SATIKAS is a remote sensing tool that uses satellite data collected by the European Space Agency to determine whether a field has been mowed or not. The detection of mowing is conducted across Estonia: all fields indicated in applications for area payment that can be mowed are analysed. SATIKAS uses the field boundaries entered through e-ARIB. Changes in the biomass of the fields over time serve as an indication of mowing. Meteorological data from meteorological radars in Sürgavere and Harku as well as from weather stations are used to refine the analysis. The mowing analysis is performed at the pixel level within the field boundary, after which the resulting data are averaged across the entire field. SATIKAS does not keep the data only to itself. Both ARIB inspectors and the applicants of area payment can access the data. The results of the mowing analysis are published in e-ARIB and the public web map that includes a separate ‘Detection of mowing’ layer.

An automated chatbot that answered frequently asked questions during the COVID-19 emergency situation. SUVE was integrated into multiple national websites and helped to find relevant and reliable information about the spread of the coronavirus in Estonian, Russian and English.

Salme is a court hearing recording program that incorporates speech recognition software. The latter has been specifically designed to accommodate the vocabulary commonly used in court hearings. A transcription of the spoken content of the hearing is generated alongside the audio recording. The text generated through speech recognition can be edited in real-time during the court hearing and the voices of the persons involved in the hearing can be better distinguished.

SayEst is the first language learning application that focuses on teaching Estonian pronunciation, providing users feedback on their pronunciation. The application includes theoretical tutorial videos and a variety of practical exercises through which users learn to recognise phones. In addition, users can compare their pronunciation to that of native speakers and receive feedback thanks to automatic speech recognition.