AI-support for reviewing building permit applications

Today, an average of 75% of all building permit cases that come to our department are incomplete. Once a case has been submitted, it is reviewed and we will get back to the applicant if we need additional information.

There are requirements from Boverket that permit decisions must be taken within 10 weeks from the time we receive a complete case.We often need to ask for additional information on at least one occasion.

Due to the number of cases and different types of tasks, it may take some time before we have time to review the case and ask for additional documents.

This creates frustration for the applicants as they often think that the case will be reviewed immediately but 1-3 weeks later are informed that they must submit additions in order for us to start reviewing the application.After conducting interviews, we have seen that the applicant often submits documents that are deficient or completely lack the necessary information; it is mainly the sketches that have flaws.

We also understood that applicants accept that the actual processing time from complete application to decision takes time, but it creates frustration that they do not receive direct feedback if supplementation is required.

Therefore, we see a need to focus on the application process and support applicants to submit a complete application on the first try.

The most common things that are missing are ground lines, direction, scale and distance to site boundary.Reviewing incoming cases and asking for additional information consumes valuable time for our employees. Time that could have been spent on other tasks.

Our research has led us to the conclusion that an AI solution that reviews drawings before the applicant submits the application would generate a much better and faster service and help and free up time for our administrators to work on more qualified tasks.Hence, the purpose of the project is to develop an AI-based solution that reviews incoming drawings through image recognition.

The AI makes an initial assessment as to whether supplementation is required or not and give direct feedback.

Within the scope of the project, we want to start training the AI on the most common errors (ground lines, direction, scale and distance to site boundary). At the end of the project, we want to have a developed AI solution that can recognize these parameters.

As a result, we want to be able to simplify the process of applying for a building permit by supporting applicants in submitting complete documents.