Your browser doesn't support javascript. This means that the content or functionality of our website will be limited or unavailable. If you need more information about Vinnova, please contact us.

AI-based dialogue support for public procurement for more relevant requirements

Reference number
Coordinator Borås Kommun - Koncerninköp
Funding from Vinnova SEK 500 000
Project duration December 2020 - February 2021
Status Completed
Venture AI - Competence, ability and application
Call Start your AI-journey for public organizations - autumn 2020

Important results from the project

** Denna text är maskinöversatt ** The project aimed to create tools for Borås municipality to increase the number of tenders and to increase the number of smaller companies as bidders. The result is a tool that helps to set the right CPV, which is central to getting more bids. The tool has also been able to identify potential bidders, even those who have never delivered to the public sector, so that they can be made aware of the possibility of submitting tenders.

Expected long term effects

The service developed will be produced in collaboration between the City of Borås and PublicInsight AB. This means that it will be available to all public organizations and authorities in Sweden. The expected effects of this are partly that the number of tenders per procurement increases, which leads to higher quality at a lower price for public organizations. Secondly, the number of companies that deliver to the public sector is increasing, which contributes to more efficient public activities and the opportunity for especially smaller companies to increase their market.

Approach and implementation

Data on most public purchases from 2019, complete data for 120,000 procurements and all public information about Swedish companies were provided by PublicInsight AB purchased through its partner Dagens Samhälle Insikt AB. The City of Borås collected its own data on all purchases for 2019. A first round of tests with wurdovec, doctovec, tensorflow and monkeylearn were performed. These gave substandard results as the procurement texts are very similar to each other. Based on experience from Lars Albinsson, a separate algorithm was created instead that could identify matchable data.

The project description has been provided by the project members themselves and the text has not been looked at by our editors.

Last updated 16 April 2021

Reference number 2020-04062