Having this topic as a starting point, Jacob Verona, a student in “Industrial Management” at Linköping University, conducted research and analysed the current methods and processes employed at Jobtech during the summer of 2023.

During the research process, the Jobtech Unit aimed at evaluating the existing prototypes in order to determine the value of automating SSYK classification for job advertisements on the Swedish National Job Board (Platsbanken), and its potential impact both for Jobtech and society at large.

Jonas Södergren, Technical Lead and Programmer at JobTech Development, provides insights into the survey’s background:

"The Jobtech Unit has implemented and evaluated automatic job ad classification since 2021. The classification has enabled the publication of a greater number of job advertisements on Platsbanken within a tab labelled “external websites”. Displaying approximately 30% more ads has created added value for job seekers. Furthermore, Jobtech has also showcased the possibility for effective collaboration with the private sector by using artificial intelligence (AI) technology.

The next step in the job classification process involves advancing the technology further, with several teams at Jobtech dedicated at developing a new technical component specifically focused on the labour market classification. Jobtech's approach is to involve multiple stakeholders early in the process. In this case, we have invited Jacob to assess and write a report on the potential of this classification."

To achieve the goal, the report involved several stages. Jacob describes the process:

"Accessing AI accuracy in comparison to human judgment was challenging due to a significant margin of error."

In the first sample, the human accuracy was assessed by conducting a test to evaluate how effectively individuals can assign the correct SSYK code to a job advertisement. The result showed that human precision was approximately 78%. Furthermore, an examination of Jobtech’s current SSYK classification algorithms and measuring their precision was conducted, which demonstrated that AI was the most effective algorithm with an accuracy of 86%. To further investigate AI’s performance, a second sample was carried out, revealing an accuracy of approximately 97%.

The report’s findings (in Swedish) state that complete automation of SSYK classification at JobTech is recommended. AI demonstrated significantly greater accuracy compared to a manual classification and has the potential to significantly improve labour market data.

Jacob provides a summary:

"Previously, Jobtech lacked quantifiable data on AI accuracy; the unit had never measured the extent of human misclassification of SSYK. Understanding human misclassifications allowed me to establish a new metric of AI accuracy."

Jacob continues to share key findings in this business case:

"The data regarding AI's accuracy allowed for the assessment of potential cost savings for the society. With the understanding that AI outperformed humans significantly, I could estimate potential societal cost savings. The question arises: should automated AI replace manual processes? One notable advantage of AI is its capacity to provide decision-makers with much more accurate labour market statistics, serving as a basis for the development of educational programs and the identification of areas of labour shortage."

Additional benefits with regards to the automation can save society more than 3,5 million Swedish kronor per year by reducing the time and resources allocated on manual job ad classification. An improvement in the work environment is also observed as automation support relieving HR personnel of their monotonous tasks. Jacob emphasizes the importance of a thorough assessment of potential drawbacks and expenses, such as algorithm management, to weigh the pros and cons effectively.

Based on these insights, a conclusion has been drawn that automating the SSYK classification process with the help of AI technology, has the potential to significantly improve the Swedish labour market. By partnering with AI, Sweden can strengthen its position in the job market and provide better service to both employers and job seekers.