Since 1999 during the annual statistical inventory of Estonian forests (SMI), field measurements take place on 5600 sample plots every year. The data provides statistical assessments of forest resources, which are the basis for national reporting and the shaping of forestry policy. The aim of the project is to develop a suitable methodology for Estonian SMI for the use of multispectral space images and aeroplane laser scanning data, to obtain assessments of forest resources as maps. The project searches for solutions for the preparation of assessment maps concerning species composition, height and stem volume of forest stands, using accurate measurement data of trees growing on sample plots as an input for the models. The results can be used for monitoring forests and planning economic decisions on the forest stand level.

A chatbot capable of independently responding to customer enquiries about e-residency.

Eia (Electronic Investment Advisor), providing e-consultations, has helped more than 5,000 potential investors since 2019 and contributed to the fact that 139 million euros have been invested in Estonia.

A chatbot developed for the Estonian Expo2020 pavilion which was built in only one language but could automatically speak many others. The chatbot could respond to topics in different languages and introduce Estonia to everyone in their mother tongue.

A prototype and conceptual solutions have been developed to launch an early warning service in Estonia.

Tagging of incoming mail. The objective of the project was the detailed analysis of the development of the technical solution for the enquiry tagger of the Consumer Protection and Technical Regulatory Authority based on the existing data set (enquiries from 2018�2021), the development of the technical solution and its integration with the email system of the Consumer Protection and Technical Regulatory Authority, and staff training.

Automated analytics tools for the Estonian language. The EstNLTK library allows for the detection of word and sentence (or clause) boundaries, lemmatisation, part of speech determination and morphological analyses and syntheses. It also recognises name entities and time expressions, providing semantic information for the latter. In addition to the existing functionalities, the library is continuously being developed. Guidelines are currently being developed so that non-programmers can also process Estonian-language texts.

Mass digitisation and image recognition of Estonia's photographic heritage

To determine the most appropriate methodology for mapping flood areas, the suitability of data from Sentinel-1 and Sentinel-2 satellites was assessed under different conditions. The necessary pre-processing steps were described and the most suitable vegetation index was selected for characterising open water. Based on the developed methodology for mapping flooded areas, a prototype for data processing was created which enables the operational mapping of flooded areas using data from Sentinel-1 and Sentinel-2.

The National Heritage Board is developing a new prototype of the AI kratt Folli for describing digital objects, the purpose of which is to move closer towards automating the description of objects, enhance data quality, allow for the staff of memory institutions to use their time more efficiently and improve public access to collections.

Using data from last year's treatment invoices of risk patients, the patients requiring avoidable hospitalisation in the next year are projected. Risk patients are chronically ill persons who are at higher risk of health problems. In addition, the model generates a risk score for hospitalisation, which can be used to rank patients.

The system records all the discussions held in the session hall of the Riigikogu as audio files, which are then transcribed by a speech recognition system. After that, HANS is assisted by an editor who reviews the text and publishes it on the Riigikogu website.