pathdoc - stock.adobe.com

Swedish municipalities test AI to drive efficiencies and cost savings

Research in Sweden measures the benefits of using artificial intelligence in local government operations

Sweden’s leading municipalities are using a research project led by the University of Lund (UoL) to test the use of artificial intelligence (AI) in the delivery of basic services. An increasing number of municipalities, grappling with the constraints of tightening annual budgets, are turning to digital solutions, including AI, to deliver cost savings across a wider range of administrative tasks.

Preliminary results from the UoL’s research programme report, Implementing digital automation in the management of supply and support, has reinforced software robots’ superiority over humans in performing a broad range of rudimentary administrative tasks in social services. Task areas tested included general data collection on clients’ income and expenses.

The AI research project, which partnered with Sweden’s 16 largest urban and regional municipalities, is funded by the National Union for Professionals (Akademikerförbundet-SSR). With more than 72,000 members, the union is Sweden’s leading association for university graduates working in public administration, human resource management and the social services domain. 

The UoL study concluded that AI robots are capable of operating at twice the rate of efficiency as human operatives doing the same work, said project coordinator Lupita Svensson.

“The use of robots in our programme has been so effective that their universal use by municipalities is now inevitable. An increasing number of municipalities will employ robots to handle social security services and support functions,” she said. “The project collaborated with the country’s 16 biggest municipalities, so it’s realistic to expect, based on the current rate of progress and development, that all municipalities in Sweden will be using robots to replace certain primary administrative tasks in two to three years.”

Workers distrust robots

Although the general introduction of robots to the municipal workplace has been positive, initial distrust of the AI technology by local authority “administrators” frequently resulted in lower efficiency gains as “humans second-guessed robots on data collection accuracy”, said Svensson.

“Robots have the ability to double work capacity, but we noticed that some municipalities initially had issues maximising efficiency gains,” she added. “Some administrators have found it difficult to trust their new software robotic colleagues. In many cases, administrators routinely double-checked the robot’s work based on accuracy and safety concerns. They wanted to be sure if applicants for certain services had understood and submitted the correct information. There was uncertainty about the technology and if robots actually work. Such constant double-checking risks losing entire efficiency gains from AI support.”

“The use of robots in our programme has been so effective that their universal use by municipalities is now inevitable”
Lupita Svensson, University of Lund

The UoL study anticipated a measured improvement in the trust values between human administrators and “AI co-workers”.

Familiarity with the technology is expected to lead to enhanced confidence in the employment of robots and a lower level of apprehension regarding AI’s ability to replace routine customer-based administrative tasks. Advances in AI technology, particularly the development of algorithms better able to compensate for human judgement skills, are expected to drive interest from municipal authorities to invest more actively in AI technology to replace the core analogue systems used to run the IT admin side of their administrations.

Significant advances in human trust emerged during test-mode AI trials at Landskrona municipality. These revealed that while robots are liable to make simple errors in calculations that may require the implementation of double-checking protocols, AI technology comprises a high degree of self-propelled efficiency potential, said Emira Talic, the head of financial support and family services administration at Landskrona.

“The trials we ran showed that the robot used for basic administrative tasks became self-propelled after just seven months of testing. The robot now handles over 70% of all routine applications each month. In test mode, the robot was tasked to handle all types of cases. In situations where it failed, improved algorithms were developed and then tested again. Once corrected, we saw an immediate improvement in the robot’s overall performance,” said Talic.

Seeking efficiency gains

Achieving greater overall efficiency in tasks output and services delivery remains the chief focal point for municipalities in their early-stage use of AI and robot assistants. With labour costs rising, municipalities are looking to digital automation solutions that have the capability to strengthen operating efficiencies across the entire services spectrum. Most (95%) of Sweden’s local authorities are currently running cost-reduction programmes that will inevitably lead to the phased replacement of staff functions by AI and digital co-workers. 

Of Sweden’s 290 municipalities, just 95 provided well-developed e-services to local authority clients at the end of 2019. Of this number, just 20 local authorities operate e-services that use digital automation functions to manage supply and support tasks.

For municipalities, the transition path to the broader use of AI has entailed root-and-branch reviews to assess their future resource allocation needs. This is being done in parallel with technical and cost analyses to determine how digital automation programmes can be best employed to optimise the delivery of local authority services at not just a lower cost, but with built-in savings to annual budgets.

Sweden’s municipalities are displaying rising interest in using digital automation solutions to deal with repetitive large-volume tasks that are linked to access to well-structured data. Increasing digitisation, conducted through dedicated robotic process automation (RPA) projects, continues to change working methods and processes. In this context, manual handling is being replaced by machines or technology in a working environment where information is becoming more digitised.

Robots in action

Against this backdrop, municipalities are showing a higher appetite for innovative digital automation solutions. The municipality of Södertälje scaled up its RPA technology capability in 2019 with the launch of the Ragnhild robot, which was tasked with performing repetitive manual and payroll functions.

The introduction of Ragnhild clearly demonstrated the potential of AI to both improve efficiency and better manage costs, according to Maria Dahl Togerson, Södertälje’s communications and digitalisation manager.

“Automation and robots are the future. They will take care of the boring routine work”
Maria Dahl Togerson, Södertälje

“Ragnhild takes care of repetitive tasks and releases staff to perform more value-added functions like customer services and resource management. Our decision to use the robot was driven by an absolute need for greater efficiency in a changing digital working environment. Automation and robots are the future. They will take care of the boring routine work,” said Dahl.

Innovation isn’t restricted to Sweden’s largest local authorities. In 2019, the minnow-sized Upplands-Bro municipality became the first user of the social interview robot Tengai to support its recruitment operations. The Tengai robot was developed and supplied by the Stockholm-headquartered AI and staffing consultancy partners Furhat Robotics and TNG Group.

“AI is very smart technology,” said Karl Öhlander, Upplands-Bro’s local government director. “The Tengai robot has revolutionised how we recruit. It accelerates the selection and hiring processes. This saves the municipality time and money, the two most critical resources we have.”

Read more about artificial intelligence in the Nordics

Read more on Enterprise software

CIO
Security
Networking
Data Center
Data Management
Close