Helsinki wants to be the most digitalised city in the world. In order to achieve this goal, the city staff conduct, among other things, quick experiments to find out how artificial intelligence and software robotics could be utilised in city services and employees’ work.
In December 2021, the city completed its third quick experiment campaign, which tested nine ideas from city employees regarding the use of artificial intelligence and software robotics:
- utilising artificial intelligence as a virtual walking tour guide
- utilising a software robot in claiming state compensation and processing interpreter invoices
- processing applications for third grade and the related pupil admission
- utilising artificial intelligence in the orientation of new employees
- timely preparation of vaccines and their delivery to vaccination points
- utilising artificial intelligence in participatory budgeting for young people
- utilising artificial intelligence in analysing responses in the Kerrokantasi (Voice your opinion) service
- utilising a software robot in calculating the value of city contracts
- utilising a software robot in the automation of student welfare statistics
A working prototype of a walking tour guide
Aiming to develop a virtual walking tour guide assisted by artificial intelligence, the purpose of the experiment was to guide people to interesting walking routes with the help of mobile phones, also providing information about cultural and historical destinations along the way. Destinations were created by combining various data materials in a new way, creating various themes on the basis of the statistical analysis of keywords. The user can choose the desired themes, the desired length or duration of the route and the starting point, and the service creates a walking tour. The destinations and their details can be viewed in advance. The experiment resulted in a working prototype of a walking tour guide with cultural history materials. According to the experiment, a service like this is worthwhile to use, and even on the basis of quite limited experimental materials, it led testers to new places in their neighbourhoods. In addition to cultural history materials, the service could be based on a wide variety of materials.
“The feedback from the testers of the walking tour guide was positive, even though it was a rather rough prototype. The basic idea works, and we hope to find the right partner and resources for the further development of the service. The walking tour guide still requires work before it can be used to create a full-scale service, but the idea has now been tested and found to work, and the building blocks are there for a more continuous service – we just need a home for it,” says Saska Lohi, Project Manager in the Urban Environment Division.
“During the experiment, I learned that we have highly diverse cultural history information but it’s often unfortunately hard to find. The experiment provided a new way of studying this information and using it to learn about history and Helsinki. When testing the experiment, I also ended up in new places in familiar neighbourhoods and got to know previously unknown destinations, which brightened my daily life during the pandemic,” Lohi continues.
Other experiments tested how artificial intelligence can be utilised in employment services to support the orientation of new employees, in participatory Ruuti budgeting to survey the wishes of young people and in classifying and analysing responses in the Kerrokantasi (Voice your opinion) service.
Software robot to assist with laborious routine tasks
The software robotics experiments aimed in particular to find ideas to facilitate manual or otherwise laborious routine tasks.
The experiments sought answers to whether a software robot can be used as an aid in claiming state compensation and processing interpreter invoices or calculating the value of contracts, and whether it is possible to automate pupil admission for weighted-curriculum education in the third grade or the extraction of statistical information in student welfare.
“The most important lesson of our experiment was that it is possible to check applications for state compensation on interpreter invoices in the future by using software robotics, which at its best could reduce manual work for employees by 0.7 person-years. This would make work more meaningful for staff, since less desirable routine tasks could be moved to the robot,” says Roope Saarinen, a computing expert from the Social Services and Health Care Division’s team that tested software robotics in claiming state compensation.
Saarinen says that the software robotics is currently being tested and minor changes are still being made before it is transferred to production. The aim is to get the robot to check invoices eligible for state compensation as soon as possible.
Ideas for timely preparation of COVID-19 vaccines
One of the experiments dealt with the timely preparation of COVID-19 vaccines and their delivery to vaccination points. At first, the experiment was to be carried out using software robotics, but eventually it was decided to use more conventional application development tools. The experiment also applied the lean development method, which helped to identify areas of improvement in the existing vaccine distribution process.
Observations relating to the timely preparation of COVID-19 vaccines were made during work at the Messukeskus vaccination point. For example, the Pfizer-Biontech vaccine must be used within six hours of being prepared and, to ensure that the vaccinations go smoothly, recipients cannot be made to wait for the preparation of vaccines. In the experiment, it was decided to create the RoTi application, which enables the timely preparation and delivery of vaccines. The application reduced the need for 1–4 people to run around the corridors of Messukeskus, collecting orders and delivering vaccinations during the last hour of operations.
Find out more:
Picture: City of Helsinki / Kuvatoimisto Kuvio Oy