(L-to-R) Fredrik Larsen, Bjol R. Frenckenberger, Lisa Z. Molbech (Source: MIR Insight)
During the 61st
Dornbirn Global Fibre Congress in September 2022, in Dornbirn/Austria, 15 young companies and start-ups presented their innovations. One of them was MIR Insight AS, Oslo/Norway.
MIR Insight provides advanced ML-decision support and forecasting on chemicals and material demand. The company was established in late 2021 with the aim to empower domain experts with explainable forecasts of future product demand based on global market activity. The method and model were tested with SMEs and multinationals in Europe and generated 90%+ forecasting accuracy on 4 months forecasts, with access to 500 million data series, and covering 40+ countries. This methodology and ML-model generate a 50% better forecasting output than traditional forecasting techniques.
MIR Insight is a multinational and multidisciplinary team of founders, with the relevant industry experience as well as scientific and technological expertise.
»The fact that we see and understand the world differently enables us to push the boundaries of innovation in a useful and creative way. Our team met during an innovation program and decided to found the company together as we combine what’s needed to understand this complex issue from diverse angles. «
Dr. Bjol R. Frenkenberger, CEO of MIR Insight
The question was: To what extent can companies be provided with more accurate insight for the upcoming months or years in terms of demand that is beyond what already know? How can better accuracy be achieved than traditional ways of forecasting and planning? And lastly – how, on top of that, can the user be empowered with new knowledge and learnings – most importantly the ability to explain.
MIR Insight has continuously revised its product, adjusted the offering based on customer needs and tested with different companies to create a superior methodology and predictive model using advanced machine learning techniques – artificial intelligence.
Unlike traditional black box ML algorithms which are hard to explain and difficult to understand even by the best domain experts, MIR Insight provides transparent ML models that produce understandable results and discover overlooked information on a continuous basis – to enable different experts in their respective domains. The customers can understand how and why their product demand has been affected by market trends throughout time and draw their own conclusions.
It is a paradox, as our society has become even more uncertain and complex on the one hand, while on the other hand there has never been so much information available to evaluate such complexity. So, one could ask, why do we still find it so hard to explain or foresee certain events or situations in a timely manner?
The company’s answer is: Too much information is not necessarily beneficial for precise decision making, it is about understanding and acquiring the correct information relevant for your needs on a continuously updated basis. This is essentially about acknowledging the fact that every product has a unique supply chain with their unique sets of drivers.
What is the challenge
Forecasting and planning problems cost companies 3%+ of their annual profits. The effect of planning problems is not only significant for profits but can also lead to reputational damage as a trusted and reliable supplier and provider, and affect the companies’ CO2 footprint. Supply chain disruptions have been constantly in the headlines since Covid-19, and a chain of new events and reactions are continuously being witnessed unfolding, which threaten to affect the value chain. Today, it seems to us that companies want to keep up with fast moving and ever-changing events and that is why the time of static forecasting approaches its end. The biggest threat to good decisions in an uncertain world is an overly strong reliance on guesswork, assumptions and not at least on the biases that stem from experience and familiarity. The world is heading towards large-scale change, which most people have never been exposed to in their professional lives. This will lead to even greater explanation difficulties.
The MIR Insight team Bjol R. Frenkenberger
, through his PhD at the University of Oxford, has gained deep understanding on how uncertainty affects decision making in organizations. He combines this knowledge with his experience of more than 6 years in AI and data analytics start-ups to bring innovative solutions to MIR’s customers. Among other positions, Bjol advanced the international expansion of start-up businesses as Global Business Manager and Business Development Lead in Fuller, Inc. and FiNC Technologies, 2 leading Japanese technology companies.
Lisa Z. Mobech
, COO, worked with multiple supply chains and industries from insurance and risk management to raw materials and chemical distribution. 7 years as an intermediary between suppliers and customer’s needs, within B2B sales and product development. She has been through the many events affecting the global supply chains pre- and post-Covid, and after witnessing the many “invisible” inefficiencies and observing the difficulties of explaining sudden events inside and outside of organizations, Mobech questioned why organizations still rely heavily on past experiences as well as information from customers and suppliers when making decisions related to forecasting and planning.
»Especially when we know how complex and unique each supply chain is and how differently affected, they are by external influence - we still tend to use the same traditional approach when making projections - there are smarter and better way to solve this. «
Lisa Z. Mobech, COO, MIR Insight
, CTO, a chemical engineer, worked more than 7 years as an environmental engineer in REC, a leading solar wafer producer. During his time there, he oversaw the implementation of environmental monitoring and LCA. He then began to work as a full-stack software developer and was active as a team lead for large scale B2B projects. During his time in Greenbird as a senior developer, Larsen was particularly focused on data integration.