Case: Experimentation accelerator
Experimentation accelerator is used to create e.g. intellectual recommender for cultural and leisure services.
Experimentation accelerator speeds up innovation
The City of Helsinki develops internal innovation activities with an experimentation accelerator. The accelerator project was launched in the autumn of 2019 with a campaign that encouraged employees to brainstorm and propose agile experiments of short duration utilizing artificial intelligence.
“Our key question with the experimentation accelerator is how to support innovation activities inside the City organization,” says programme manager Ville Meloni at the City of Helsinki office that steers the City’s digital transformation.
“With the help of the accelerator, we seek to create a mechanism to support agile experimentation utilizing digitalization. Our goal with the mechanism is to distribute lessons learnt on experimentation, as well as to scale workable solutions in the organization.”
The experimentation accelerator has grown from design thinking. “Experimental development is one method of design, in which common understanding of the most workable solutions for users is sought one small step at a time,” Meloni explains.
Design thinking is implemented in the project by co-creation involving employees, users, and enterprises providing expert services. An important measure of success is how the lessons learned are documented and shared with the organization. The entire process is made transparent with active communication.
“We’re dealing with an organizational culture change,” Meloni asserts.
Experimentation accelerator launched with AI experiments
The experimentation accelerator was launched on the basis of interview research conducted by the Demos Helsinki think tank.
The interview research discovered several hurdles to experimentation routines inside the City of Helsinki: lack of time, resources, and skills in experimentation. There are also many challenges related to procurement and the scaling of experiments.
The outcome was the first stage of the experimentation accelerator.
The platform of the accelerator was the City’s internal ideation channel Ideapaahtimo (idea roastery). Over one month, any City of Helsinki employee could propose ideas for artificial intelligence (AI) experiments.
“We built a low-barrier campaign. You didn’t need to be an AI professional. The City provided an AI course and teaching videos for anyone interested in the topic,” Meloni says.
“The participants were supported with their ideas on the Teams channel and in experimentation clinics. They were guided in the clinics by about ten enterprises specializing in AI. The campaign was presented online, and it was discussed face to face in a breakfast meeting.”
“We received 31 proposals. This signifies clear interest in such experimentation activities that accrue knowledge in the possibilities of AI in the development of City operations and services.”
The seven best ideas were selected for further development and granted 10,000 euros each for implementation.
A recommender for cultural and leisure services
Sari Lehikoinen (City of Helsinki), Jaakko Matomäki (Deloitte) and Ari Tolonen (Digitalist) in an experimentation accelerator workshop.
Communications specialist Sari Lehikoinen at the City of Helsinki Culture and Leisure Division envisioned how to utilize AI to attract new customers and to improve services.
She had been acquainted with YouTube’s recommendation algorithm as part of earlier AI training and was thinking, “We could do something similar, cross-marketing our cultural and leisure services on our website. For example, when customers read about a City Museum exhibition on our site, they could receive recommendations for books and concerts of the same theme.”
Lehikoinen submitted a project proposal named Löytö (discovery) to the experimentation accelerator. Löytö was selected for implementation.
“Löytö can be implemented on a small scale, but it could be scaled into a large concept for the entire City,” she says, emphasizing the benefits of the cross-marketing tool. “Löytö is a pilot and pioneer.”
More efficient fire inspections
Riku Leppänen (Helsinki City Rescue Department) and Teemu Ruohonen (Rojekto) discussing how AI could be used to improve fire inspections.
Fire inspector Markus Latva-aho of the Helsinki City Rescue Department recognized a clear drawback in the department’s mandatory fire inspections: inspection targets are selected manually from a huge quantity and assigned to fire inspectors according to the supervisor’s discretion without any automated system.
There has to be a better solution, Latva-aho thought. He had studied innovative technology and visions for the future. “If AI could be used to assign fire inspections, they could be targeted more efficiently, and they could be more goal-oriented and based on real risks,” he envisioned.
But how to realize this AI concept? Nothing like that had ever been done before anywhere else. The answer came from the experimentation accelerator.
“Without the opportunity offered by the experimentation accelerator, I doubt that I had ever found a channel to realize the idea,” Latva-aho comments.
“The accelerator made it possible to receive funding for my idea and to launch research and development. The City gains knowledge that can be applied in other connections.”
Operation evolves and becomes established
The AI campaign experiments were launched in December 2019, and the work continues until the beginning of March 2020. The developers meet regularly in co-creation meetings.
“The main goal of the first round of experimentation is learning,” Meloni emphasizes.
“We want to discover new opportunities offered by AI. At the same time, we want to develop the experimentation accelerator to better serve the City’s innovation and experimentation activities.”
Text by Johanna Lemola
Photos by Vesa Laitinen, Tero Lahti, City of Helsinki