How AI can help those out of work

We too often hear stories about people getting lost in the system. Contact with municipalities, job centers and The Social Insurance Agency can be tough. Especially when you have a situation that involves more than one of these organisations. And even more tricky if Swedish is not your first language. We talked to Ruth Lund, Project Manager at Samordningsförbundet Centrala Östergötland, about their project where AI will help those out of work by making the process easier.

Ruth Lund, Project manager at Samordningsförbundet Centrala Östergötland.

Samordningsförbundet Centrala Östergötland is a coalition of organisations dedicated to make communication between The Job Center, the Swedish Social Insurance Agency, Region Östergötland and the Municipalities of Linköping, Åtvidaberg and Kinda easier.

–       For a person in need of service from just one of these parts it is fairly easy to manage. But if you’re in need of several of them it gets quite confusing and takes a long time to work through all the steps. This leads to many people getting stuck in the system and it is for these groups of people that we exist, explains Ruth.

A new knowledge centre

In contrast to other similar organisations Ruth and her team work a lot with innovation and are for example developing a National Knowledge Centre called Welfare in Collaboration where they work together with the Universities of Linköping, Karlstad, Reykjavik and Stirling.

–        We stay close to the academic world and collaborating with researchers to find new things to try. We look at the things that don’t work and try out new ideas and experiment to find better ways.

One of the projects that is already up and running is called Insatskatalogen, a catalogue that collects all programs and initiatives for people out of work. It makes it easier for government administrators to provide good service and point people in the right direction.

–        It started as a regional project but rather soon it was clear that this need existed nationally. Right now, we are about 60 municipalities all over Sweden using the tool and it covers about half of the Swedish population.

Guidance for people lost in the system

As mentioned, this tool is mostly for the employees that help the target group. The natural next step would be to examine how to help the actual target group with a digital tool, and this is where the AI comes into the picture.

–        Our mission statement is that “We provide guidance for people lost in the system in a useful platform with a tailored voice, helping them feel empowered to change their lives.” And what we hope to create is support for different people in need.

This project is called the Welfare Guide (Välfärdsguiden). It is still a prototype but the goal is that it will be a tool for people who want to start working. It can be people who are longtime sick, have a disability or are socially isolated.

–        The situation right now for this target group is that there are loads of useful information online for them. But it is really hard to navigate, and it is also hard to know who to turn to for help. The platform that we are working on will be a bit like 1177, but for welfare.

Not just digital- but smart

The first step is to create a site where all useful information is gathered to make it easier for the target group to find their way through the welfare system. But hopefully this will not be the final product, the goal is to make the platform smart as well.

–        We want to connect Artificial Intelligence to this site, to be able to tailor a guide for every specific user. We are cooperating with several different partners at the moment, both from our different networks such as AI Sweden but also with a similar project in Iceland that is in use right now.

The goal is a smart guide that will calculate what kind of activities I should do, who I should contact and in what order depending on my specific situation. The optimal route for me will be based on others that have had the same prerequisites. It will also be able to give an approximate timeframe for when you will be able to get back to work.

The similar tool that is used on Iceland is research based and working today. The difference is that it is not used by the target group themselves but by employed administrators.

–        The challenge is to take the things that work in Iceland and adjust them to our Swedish systems, which are very different. 

A push in the right direction

One obstacle with the project is the laws surrounding Personal Data. For the Welfare Guide to be able to function there is a need to access and save a lot of data that is protected under the current system.

–        We believe that being able to share data like this would create great value for the target group. In healthcare we see it as a given that information about ourselves is available to the people treating us, but it is not so clear in welfare.

Ruth hopes that projects like this will give a push in the right direction to the legislators of Sweden since a change has to be made to make the smart version of the Welfare Guide a reality.

Another aspect of this is explored in a VINNOVA funded project that Ruth manages together with Petter Falk from Karlstad University, “Att se sig själv i koden” (See Yourself in the Code) which is a collaboration with Linköping Municipality during 2021.

–        Within municipalities and similar organisations the digital tools used are often picked out by technical experts and not by the employees who actually will be using them. So in this project we want to educate the employees in digital tools and different kinds of AI and show them that they are the actual experts since they are the ones working where the data originates. 

Samordningsförbundet Centrala Östergötland has Linköping Municipality among its member organisations – a partner to AI Sweden. If you want more information about AI Sweden please contact Node Manager Niclas Fock.

Europeiska unionen Europeiska regionala utvecklingsfonden
Europeiska unionen Europeiska regionala utvecklingsfonden