In 2023, generative Artificial Intelligence (AI) gained significant attention, with a wider audience being introduced to language models for the first time. St1’s Digital Process Automation team within the Business Technology unit seized the momentum, and developed St1’s own in-house AI assistant, Pluto.
As the hype around generative AI grew, the need to provide employees with a safe and easy way to get familiar with utilising large language models (LLM) became clear. “The number of solutions that can utilise artificial intelligence is growing rapidly, which is why it is important for everyone to have a basic understanding of AI,” says Onni Kultalahti, Service Owner of Digital Process Automation at St1’s Business Technology unit.
Onni began testing the new generative AI’s capabilities with a team. They quickly developed the first version of St1’s own application, Pluto, which utilises a large language model (LLM). The application was rolled out for all employees in autumn 2023, with a clear goal of engaging them on the potential of this emerging technology.
“Before gaining access to the tool, each employee needed to undergo training, where they learned the basics of AI and how to utilise generative AI responsibly” Onni explains. Training sessions were offered face-to-face and online in various locations.
“The interest towards AI and Pluto was high, with a large number of employees participating in the training and subsequently using Pluto,” says Onni.
Benefits of in-house AI
The Digital Process Automation team is aware of the constant advances in generative AI applications more broadly. However, Onni sees several unique benefits in having an in-house AI assistant tool. First, the rollout of an in-house tool was able to capture the attention of employees and help educate them on the uses of generative AI and AI in general.
Secondly, maximum data security was achieved, up to the point that employees can even use Pluto with confidential information. Because of this, we can use it for actual business purposes. “All the elements and knowledge regarding the application are within our control. Pluto is a closed solution and operates within our own cloud environment, minimising the risk of data leakage,” Onni explains.
Onni thinks that building Pluto was also a valuable learning experience for the team: “the best way to learn is to get your hands dirty. Now we understand better how certain AI solutions are built and how to best utilise them.” This brings long-term benefits, as it will be easier to integrate new generative AI solutions for various business purposes.
Pluto 2.0 is on the way
Since launching the first version of Pluto, the Digital Process Automation team has already made an upgrade.
“Pluto 2.0 was developed as a result of listening to user feedback. The first version is fairly simple, and the upgrade brings significant user experience improvements.” Pluto 2.0 has a more modern user interface, possibility for real dialogue, and new functionalities, such as the option to connect different St1 data sources to the tool.
At the end of 2023, the technical capabilities for Pluto 2.0 were ready, and it was being tested by a small pilot group.
An entrepreneurial approach to innovation
For Onni, an important aspect of working at St1 is the approach to exploring innovative solutions and having the courage to try them out. “If you come up with an idea with potential, we dive straight in and start testing,” he adds. The way of working in the Business Technology unit is agile with a flat hierarchy. This gives employees the responsibility and the opportunity to pursue new ideas that can contribute to the success of the company.
Onni is happy with the support and encouragement he received from team leadership in the early stages of the development. “Back then, generative AI was a fairly new concept to us. However, we decided to take the leap, and the result was a success.” Onni believes this entrepreneurial approach describes St1’s company culture in general.