Andritz – AI assistant

Designing a human centric AI assistant to maximize value beyond technology

Artificial Intelligence technology has taken big steps forward in recent years and can help us in many ways in our daily tasks and routines For example, AI systems are good at doing repetitive tasks that humans may find tiring or they can help collect information from several places and remember important things that humans easily forget.

Utilizing artificial intelligence to minimise routine tasks

Artificial Intelligence technology has taken big steps forward in recent years and can help us in many ways in our daily tasks and routines For example, AI systems are good at doing repetitive tasks that humans may find tiring or they can help collect information from several places and remember important things that humans easily forget.

However, when introducing new technology that affects work routines of the users, caution must be taken to design a tool that is easy to adopt and serves the best interest of the users. The question is, how can we design an AI assistant from a human centered perspective so that it makes human work more enjoyable and efficient?

Background

A significant part of the Andritz spare part sales team’s time and effort was spent on routine tasks investigating details related to customers’ inbound service requests. The initial hypothesis was that AI could help identify the key details faster and more accurately than humans.

The hypothesis had been validated by a proof of concept project by Gofore where it was tested how an AI assistant could automatically identify and gather the required information from customer emails. However, it was yet to discover how the sales team felt about this type of solution and what were the biggest bottlenecks in their work process

Designing the assistant by co-creating with users

Gofore conducted a service design project to better understand the work of salespeople and recognize how the technical AI solution could be utilized to make it truly valuable for the team. The sales team was involved in a co-creation process to identify the optimal way to introduce the new assistant to support and boost their daily work.

A multi faceted team with experts in user research, UI design and data science was assembled to complete this task. We conducted user interviews and observations to understand how the overall process from request to delivery was completed in the current IT system. Emphasis was put on understanding how they feel about the system, how does it work in their favor and what are the
things they struggle with.

In addition, we conducted a workshop together with the users where the work process was constructed on a canvas and improvement plans were clarified. The sales team was involved in the design process from start to finish, so they knew what
to expect from the new AI Assistant and were excited to welcome this type of technology in their work.

During the process it became evident that there were also other equally significant bottlenecks in the process that could be alleviated with further AI assistant features. Understanding the overall work process enabled us to design a comprehensive solution
that will bring much more value to the team and help them to get rid of repetitive tasks. It helped to add value beyond the technology!

Benefits from a human-centered approach

“Involving users actively in the design process was something I really want to continue doing in the future. They are the real experts of their daily tasks. It was great to experience their enthusiasm towards developing the AI assistant.

It was also insightful to see how Gofore experts were able to pinpoint the most important aspects that could be solved with artificial intelligence. Actually, I was amazed to see that utilizing AI like this in our work is even possible. Spare parts are an important business for us, which is why it is important to make the challenging work more efficient and reduce the amount of manual work for the experts.”

– Satu Svensson, Spare Part Manager / Spare Parts, Andritz

“I find this type of project great and rewarding, where the people whose work is being developed can take part. It was especially important to feel that I was heard. When I could explain the problematic parts of my work, they became visible and we were able to focus on those from the very beginning.

The commitment towards the results and possible changes in workflow is at a much higher level, when I could see for myself how and why these decisions were made. The project also increased awareness of my own work and pride of the versatile experience it requires.”

– Seija Manninen, Service Engineer / Spare Parts, Andritz

What did we learn during the project

“The project reminded us how critical it is to involve actual users in designing an AI assistant. Without their perspectives and stories on how they use the system in their daily tasks it would be hard to understand, what problems we are trying to solve with artificial intelligence.”

– Hanna-Riikka Sundberg, Senior Business Designer, Gofore

“It was truly inspiring to combine artificial intelligence with human centered design process . I felt many eyeopening moments during the project, when I witnessed the benefits of the utilisation of AI can have to overall user experience. We should always unite user insight with intelligent data by default Teemu.”

– Miika Mikkola, Senior UX Designer, Gofore

“The project helped us to understand how the AI assistant can be designed in a way that makes the life of the sales team
easier instead of being yet another application they must start using. We also understood that it is important to design a
positive feedback loop, so that the users find the assistant so useful, that they want to use the assistant, and the assistant learns from the users becoming more useful all the time.”

– Teemu Hirsimäki, Senior Data Scientist, Gofore

Project highlights

Involving user insight

With the end-users insight the team were able to identify the true problems the AI was going to solve.

Designing an AI assistant

The AI assistant was designed to support people and make their work easier by minimising routine like tasks.

Benefitting from a human-centric approach

With co-creative methods the end-users felt involved during the whole process. The users knew what to expect and welcomed the AI assistant with excitement.

Expertise