AI and text analytics – Piloting with public procurement data

käsi joka pitelee digitaalista rubiikinkuutiota

How can AI and text analytics help to analyse public procurement data? KEINO attended EU meeting with Commission on electronic procurement and data analytics, where it presented how this has been piloted in Finland.

KEINO supports and helps Finnish public procurement experts and authorities with the development of sustainable and innovative procurement.

KEINO's main targets are to ensure that the number of innovative and sustainable procurements in Finland increases, and that public procurement is recognised and actively used as a management tool and public sector learns from one another. As a part of these targets, it is important to also develop ways to measure, track and analyse public procurement data.

Two ways to get statistical data about innovative and sustainable procurements

KEINO is already piloting text analytics and big data of public tenders and attachments. It has launched systematic and real time way to do text analytics related to published open tenders and all the texts in tender documentation. The new technique helps in analysing the tenders and attachments in more detailed criteria level what it comes to innovations and innovation friendly processes and different aspects of sustainability. For example, about 20-30 percent of tenders represent innovative technology or includes a feature that favours innovations in the procurement process.

In addition, KEINO utilizes statistical data. In Hilma there is a national tender report, where it is possible to report in manually chosen innovative and sustainable options, which is mandatory in every public tender.

Further steps in Finland using AI in public procurement

At the moment, KEINO and Motiva are building national real time criteria-database which is connected to different procurement planning solutions in the future. In addition, Hankinta-Suomi project and KEINO are continuing their work to develop e-tools and monitoring practices and scale them up in national level.

The already piloted AI and text analytics model could be easily translated and widened to TED and EU’s national databases using developed data scraping methods and key words. During the coming year, KEINO is piloting a long term analysis of big data with massive amount of tender documents in analyzing sustainable and innovative procurement and suppliers in local and national level. In the near future, the produced big data can be combined with GBT-4 and other text models to help procurement units, firms and society.