Scm Group: Servitization, Control room, Artificial intelligence

An investigation by the publication Industriaitaliana.it on strategies and new business models tied to servitization, highlights Scm Group as a prime example among Italy's leading companies.

Alessandra Benedetti, Head of Service & Spare Parts and Digital Transformation & Business Remodeling of the Group, shares insights in a recent interview.

How is Scm Group adapting its industrial strategy to embrace the shift towards servitization-based business models? What new initiatives are being pursued?

For Scm Group, embracing servitization means not just offering value-added services that enhance performance and user experience. This is a head fake: in that we are facing a thorough change management project.

This transformation involves integrating advanced technologies while prioritizing a seamless user experience. It seems like a paradox, but we need to use complexity to make our customers’ lives easier. Specifically, Scm Group is focusing on extending the product lifecycle and enhancing customer relationships through innovative solutions:

• Monitoring and Proactive Support: Utilization of IoT and data analysis tools to monitor products and their usage in real-time, enabling proactive interventions to maintain or optimize asset performance, including energy consumption.
• Investment in Emerging Technologies: Adoption of advanced technologies including symbolic AI and Machine Learning to automate internal processes and create a shared technical knowledge management system, enabling easy access to solutions and “how to” tutorials, in a knowledge-as-a-service paradigm.

Condition monitoring and predictive maintenance: where do we stand? How is the market responding?

Last September, during the inauguration of the Rimini Technology Center, Scm Group introduced Control Room, a new concept of proactive service enabled by IoT, edge and cloud computing, AI, and BI. This platform allows us to anticipate potential issues monitoring in real time the operation of the machinery in the field while running algos in the background to catch any possible anomalies by cross referecing variables and analyzing time series.

By investing in innovation, which is not only technological but increasingly service-oriented, we aim to convey to the market the message that differentiation and the ability

to create partnerships with our customers depends on permeating technologies with meaning and value. Investing in technologies is not sufficient for innovation. Instead, we must conceive and co-create unique and divergent use cases, fostering a new capacity for active and empathetic listening to our customers. Our customers are full of ideas waiting to be developed!

What is the roadmap to successfully implement digitalization enabling servitization?

It's not just about having a vision or the ability to cleverly imitate. Developing expertise and know-how, as well as investing in the best technologies or start-ups are not enough either. What appears necessary and paramount to us is to create an internal organizational context, with strong ties to the market and the territory. This context should be capable of nurturing and fostering new forms of collaboration and new ways of conceiving and developing product and service innovations. It's about the mindset, not just the skills. It's about the team, not just reinventing the process.

Are there opportunities for development in both generative and non-generative artificial intelligence?

We're heavily investing in exploring and prototyping AI-driven use cases. Our applications range from classic statistical AI to areas like time-series, computer vision, text analysis, and human-machine interfaces.

All the above techniques are not mutually exclusive but rather complementary. The specific level of success of each technology strongly depends on the process of designing the service behind it and not vice versa.

For example, consider the challenges of applying natural language understanding in an industrial context. We recently released a proprietary solution for self-service access to knowledge that is organized and distributed (or, to put it more accurately, "dispersed") across our systems, both legacy and cloud.

Before publishing the beta version, we developed and compared two alternative proofs of concepts, both aiming for the same goal. The first one was developed by using only symbolic NLP rules and a semantic engine for data classification and extraction, while the latter was implemented with GENAI (LLM) and trained on the same "technical publications" that defined the training set for the first. User feedback and KPIs led us to choose the non-generative artificial intelligence solution, while the GENAI solution was less accurate, less precise, less secure, less repeatable, much more expensive, and less sustainable.

However, we don't intend to dismiss "general-purpose" artificial intelligence outright just because they didn't meet one challenge ideally. We believe they may be perfect for other use cases that we already have on our roadmap, where they can be better "specialized" and "contained.