User-centric Generative AI solution enhances Metso's technical support

As a frontrunner in end-to-end solutions and services for aggregates, minerals processing and metals refining, Metso wanted to improve the technical support for its end customers.   

In the event of an incident, communication between technical support and the customer is critical, so delays and fragmentation of information needed to be minimized to ensure the best customer service. Metso had identified a need for Generative Artificial Intelligence (Gen AI) capabilities in enhancing technical support and was looking for a partner whose expertise would cover not only customized solutions for generative AI, but also user insight in technically demanding environments.  

Technical support technicians had to navigate vast amounts of data in databases as well as tacit knowledge of the whereabouts of files in scattered locations. As a result, problem solving for technical support could sometimes be time-consuming, requiring a lot of background work and asking the customer clarifying questions. This made it difficult to deliver fast and high-quality customer service.

Knowit was selected as the implementation partner for its recognized holistic expertise. The project objectives, end-user needs, and AI solution ideation were refined by applying design thinking methods and the AI Design Sprint model using AI technology. Due to the rapid conceptualization, a chatbot utilizing generative AI was identified as a suitable solution to streamline the technical support team's work.   

From the very beginning of the project, the people of Metso recognized that user orientation is a prerequisite for success, especially in industrial systems. The solution's interaction with the user was designed specifically for professionals working in technically challenging environments, so all chatbot responses contain a linkable source reference to the original data, thus validating the expert's knowledge.  

As part of the project, Knowit helped with data aggregation, processing, and management. This allowed the chatbot's search function to be customized to take advantage of sources that cannot typically be retrieved from databases. The chatbot runs on Open AI GPT and Microsoft Azure technologies. 

The solution now acts as an expert search engine and data processing bot, allowing technical experts to quickly find answers to complex customer questions about Metso's equipment, installations, maintenance, and incidents. It is a tool to enhance the end-customer's work and Metso's customer service, supporting information sharing and quality assurance. The solution speeds up daily customer support work and the orientation of new employees.  

The impact of the chatbot will be reflected in the meaningfulness of the technical customer service staff's work and the efficiency of the end customer's work, and in the extension of the equipment's life cycle. When maintenance needs can be addressed quickly, downtime is reduced, and utilization improves. This allows Metso's equipment to achieve higher productivity while reducing emissions.  

An easy-to-use interface that considers the reliability of the information provided to its user allows the chatbot to be robust, fit for purpose and genuinely serves its user's needs. The chatbot is built using Infrastructure as a Code (IaaC) methods, which allows the solution to be cost-effective and quickly scaled up to future variants using the existing technology base. AI solutions require monitoring and active feedback from users, so a common code also enables agile maintenance. The expansion of the solution across the entire equipment base has already started in cooperation with Knowit .   

Metso is a frontrunner in sustainable technologies, end-to-end solutions and services for the aggregates, minerals processing and metals refining industries globally. Metso improves their customers’ energy and water efficiency, increases their productivity, and reduces environmental risks with their product and service expertise. 

What we did 

•     Customer-centric design 
•     Gen AI 
•     Software development 
•     Cloud services 
•     AI Design Sprint