A major challenge facing the job market is matching job seekers to employers. The Swedish Public Employment Service – with the help of RISE – has developed the service Spontanansökan (Spontaneous Job Application) using existing data and applied AI to improve job matching.
“We asked ourselves what we could do with all of the data we were sitting on,” says Simon Benjaminsson of the Swedish Public Employment Service. “We wanted to develop data-driven matching services; to see how far we could get using existing data.”
One initial result of this reasoning was the service Spontanansökan, which allows job seekers to obtain a picture of how likely it is that an employer in a given district and sector will be looking to recruit; so that they can send out spontaneous job applications with greater accuracy. The service is based on data from three public authorities; the Swedish Companies Registration Office, Swedish Tax Agency and the Swedish Public Employment Service itself.
“Data from he Swedish Companies Registration Office and Swedish Tax Agency offer an indication of whether the company is likely to be recruiting. Data from the Swedish Public Employment Service give an idea of what they might be looking for, if indeed they are considering recruiting,” explains Simon Benjaminsson.
The database is the key
The Swedish Public Employment Service itself is behind the Spontanansökan service, which is intended as part of the forthcoming Jobstore initiative – an umbrella for all labour-market related services. However, according to Simon Benjaminsson, in reality the service is of secondary importance.
“The database on which the service is built, along with its open APIs, is the key factor. We want to open up our data and analyses and cooperate with other parties such as public authorities, municipalities, recruiters, employment agencies or researchers. The hope is that this will lead to services that we haven’t though of ourselves.”
The Swedish Public Employment Service has itself developed the current demo version of the service with advice from RISE. RISE has also carried out data analysis for the project and developed a graph of all Swedish employers, showing their similarities.
“By using text analysis technology on organisational descriptions obtained from the Swedish Companies Registration Office, we were able to make a more detailed classification and grouping of companies,” explains Olof Görnerup, senior researcher at RISE. SNI codes (the Swedish Standard Industrial Classification) provide only a rough classification. Using this technique, we have been able to provide a considerably more detailed picture.
AI expertise is the key to success
According to Simon Benjaminsson, that the Swedish Public Employment Service chose to turn to RISE for help was a matter of finding the correct expertise.
“The right background and expertise is required in order to accomplish this kind of task. When it comes to cutting-edge technology, it is not always easy to assess which competences are needed. As I have a PhD in computer science, I knew where to find the top experts in the field,”
Spontanansökan may be the first service that the Swedish Public Employment Service has developed based on existing data and artificial intelligence but, in all likelihood, it will not be the last. They see AI as a tool for streamlining matching in line with the authority’s overall objectives.
“AI is part of the digitalisation that is underway throughout our organisation,” says Jonas Södergren, the man responsible for Jobtech at the Swedish Public Employment Service. AI should not be used for its own sake, rather only where it can offer benefits and facilitate matching job seekers with employers who are looking for personnel.”
Launched as a pilot scheme in autumn 2017, the Swedish Public Employment Service expects Spontanansökan to become more widely available in the latter part of 2018.