The City of Helsinki’s Experimentation Accelerator promotes agile experimentation with digital and AI-based solutions to support service development. CGI is involved in projects where AI is being tested, for example, in forecasting workforce needs for library services and improving information management in infrastructure construction.
The Experimentation Accelerator is a support service for City of Helsinki employees designed to facilitate agile experimentation with new ideas that use digital technologies. The accelerator assists with the planning, procurement and implementation of pilot projects. Its goal is to encourage learning through experimentation and help the city further improve its services and operations to become more customer-centric.
We’re involved in several pilot projects that explore how AI can be applied in practice across different sectors—from workforce planning to infrastructure construction.
The pilot projects demonstrate how AI helps public organizations move toward data-driven operations while freeing up resources for human-centered work through more efficient use of data and streamlined processes.
The pilot has provided significant insights into the systematic use of GPT-based AI. Carefully-designed task assignments, comprehensive documentation, systematic use of previous results and quality assurance in the interpretation of numerical and textual data have reinforced the reliability of AI-based scenario modeling.
Development Manager, Helsinki City Library
Using AI in workforce planning
The Helsinki City Library aims to modernize its workforce planning to better respond to changing operating environment and service needs. New tools are needed to support decision-making and enable more accurate forecasting and allocation of human resources.
The goal of the pilot is to develop an AI-assisted solution that supports staff allocation and forecasting by taking into account the specific characteristics of different library branches, as well as regional changes in population and service needs. At the same time, the project helps strengthen employees’ capabilities in using AI solutions as part of ongoing scenario planning.
The pilot project is building an AI model that uses multiple data sources including population forecasts and customer usage data to generate future staffing scenarios. The tool will be tested together with supervisors in library services, and the solution will continue to be developed based on feedback from users.
"The pilot project has provided valuable insights into the systematic use of GPT-based artificial intelligence. Carefully designed tasks, comprehensive documentation, systematic use of previous results and quality assurance in the interpretation of numerical and textual data have reinforced the reliability of AI-based scenario modeling. At the same time, the pilot participants’ AI expertise and understanding of its operating principles have developed significantly. The fifth version of ChatGPT, released during the pilot, enabled a significant leap forward. Without it, the project’s goals would not have been achieved as planned,” says Matti Sarmela.
Improving information management in infrastructure construction
AI also offers significant potential benefits in infrastructure construction. Currently, project-related information including instructions, requirements and documentation is often spread across different systems, making information management more difficult and increasing the risk of errors.
The goal of the pilot project is to develop an AI-assisted solution that supports project information and project management. The tool assists employees during different stages of the construction process, such as preparation, change management and decision-making during construction. Improved information management can also speed up the resolution of issues that arise during warranty periods.
The AI tool being developed searches project documents for current and relevant information while applying defined constraints. The pilot project will identify key data sources and determine the situations where AI can provide the greatest benefits. Feedback from experts will also be collected and used in further development.
AI can help identify unclear or ambiguous wording in contract documentation at an early stage, enabling timely corrections, supporting more efficient contract execution and potentially reducing ambiguities related to unexpected costs.
“The pilot project makes use of actual city infrastructure projects that are at various stages of construction. These include projects that are in the planning stage, currently underway and already completed. For these real projects, we can identify the specific stages and benefits where AI can deliver the best results. The pilot will also identify areas for improvement in the city’s various documents and operating procedures. At the same time, we’ll see where AI isn’t worth using or can’t be used,” explains Ilkka Julkunen, Project Manager at the City of Helsinki.
Both trials demonstrate how AI can support public organizations in their transition toward smarter, data-driven operations. By improving the use of data and thoughtfully automating processes, organizations can free up resources for work where human expertise remains essential.
