Between digital transformation and climate change, many companies are involved in the most detrimental process of change of the 21st century. Environmental change poses both physical and economical risks to businesses. Due to an ambitious climate policy, most economic branches already feel the pressure to put environmental measures into practice. Digitization and climate action now make the header of every company’s annual agenda, but should these two goals be approached separately, or are the incentives nevertheless intertwined?
Sustainability is profitability, indeed.
It seems that some have already found the point of intersection. The use of artificial intelligence in business processes has positive repercussions on the sustainability of the company as a whole (Schulzki-Haddouti, n.d.).
In the industry, machine learning is used for predicting events that one might want to correct, adapt or even avoid. It leads to higher efficiency and less wasted resources. According to the study by Jochen Flasbarth and Martin Vogt of the VDI Center for Resource Efficiency, 42 percent of the companies surveyed use AI-based solutions in their own production processes. Time and cost savings, as well as improved quality, seem to be the main reasons behind this. Additionally, almost all of them see an environmentally relevant potential for saving energy and materials. (VDI ZRE, 2021)
As an entrepreneur, have you already thought about it? What does your climate policy look like, and could digitization be the answer that catches two flies at once?
In the lines below, we present two application examples of AI in the industry that have an impact on resource- and cost efficiency.
Prevention is better than the cure
In order to save production costs, but also to reduce their energy consumption, several companies have incorporated machine learning into their operational processes. The study "Potenziale der schwachen künstlichen Intelligenz für die betriebliche Ressourceneffizienz” by VDI ZRE uses practical examples to analyze the effects of artificial intelligence on saving company resources.
Predictive maintenance is the result of using AI methods in production. The aim of the study was to determine if the predictive maintenance of milling machines would have an impact on steel production overall.
Since the cutting head of the milling machine experiences a lot of wear and tear, the precision of the cut suffers, which in turn has an effect on quality, resulting in more scraps. With the help of AI, the condition of the components is being monitored and the remaining life cycle of the machine is determined.
The result? Increased reliability, small deviations, and real-time detection of anomalies. The production plant has experienced less material waste through the early detection of inaccuracies. The increased precision of the cut has also eliminated additional processing, thus saving on unnecessary energy consumption and extra material. Finally, through these improvements, the company’s overall CO2 emissions have been reduced.
Another use case of predictive maintenance involves a leading manufacturer of commercial vehicles in the automotive industry. The company was struggling with recurring transmission damage in the produced vehicles and came across craftworks, in the hopes of a solution to predict such malfunctions in a timely manner. The aim was to reduce the high warranty costs and customer dissatisfaction.
What’s more, since the vehicle experiencing the damage to the gearbox was usually already in transit, higher costs and carbon dioxide volumes would arise due to the vehicle having to be transported to a service center. We took a look at the matter at hand and our Data Scientist, Tatevik Gharagyozyan, started working on a solution. The goal was simple, a solution for early detection of gearbox failures so that the manufacturer could take preventive action and avoid mishappening. For the POC, the company provided aggregated operational data over a period of ten years, which was used in the creation of a predictive maintenance model. In 60 percent of the cases, failures were predicted correctly. To improve the model accuracy, higher data quality would have improved the machine learning process. Nevertheless, the impact on the producers’ business was substantial. The manufacturer saved not only on warranty costs but also managed to therefore lower its CO2 emissions.
Looking at another industry, craftworks has undertaken another predictive maintenance application with one of Austria's largest energy providers. Heat accounts for 50 percent of the total energy consumption, and the providers supply the heating infrastructure throughout the city of Vienna. On the company’s site, specialized staff performs the inspection and maintenance work. Back then, when the project started, an automatic warning system would inform of any faults but was failing to indicate the severity of a detected problem. This meant that each warning had to be investigated and staff had to interrupt their ongoing maintenance to see what the warning was indicating. Time was being wasted. The energy supplier was looking for a solution to increase efficiency. That's how we were introduced to the matter and our Lead Data Scientist, Daniel Ressi, came in and took over the project.
Our concept was to implement a model that could predict whether a disruption would occur within a certain time window. We then implemented an additional model to detect the disruption type if the first model predicted a high probability of disruption. Such AI technologies can be used by companies to predict the occurrence of an undesired event.
This will save time and cost resources, internal processes will be optimized. Constant, operational digitization brings with it environmentally friendly results in production and maintenance, thereby reducing CO2 emissions throughout the supply chain. This brings companies closer to their sustainability goals.
Alright, so … what’s next?
How do you implement this in your company?
Well, for starters, you have to ask yourself a few questions. For a successful AI implementation, you need sufficient data, which also determines your company's level of digitization. (Deloitte Consulting GmbH — Analytics & Cognitiv et al., 2021). Making use of Big Datasets and being able to predict future events is made possible with the help of artificial intelligence. Going forward, the one question that companies ask themselves at the start of their AI journey is: what budget and time resources do I need?
What we observed in our partnerships with our clients is that, regardless of the size of your budget, it takes time to build expertise. Even if you have the financial means to build an extensive data science team overnight, there is still a learning curve for the team to share knowledge, find their strengths, and understand how to work together cohesively. In this case, bringing on board an experienced AI Solution Partner like craftworks can accelerate the initial stages of the project tremendously. With an extensive portfolio of clients across the industry, the data scientists of craftworks have already experienced all facets of the AI-related projects several times and know how to work together efficiently, right away.
After a project is set up and deployed at our partner’s site, craftworks provides its clients with knowledge transfer and guidelines for further development and maintenance of the project. Last but not least, craftworks has created navio: a platform that automates machine learning operations. The user-friendly interface of navio lets customers deploy, manage and monitor machine learning models across an organization’s entire AI landscape.
As a result, machine learning can become an integral part of your daily operation without the need of extra resources or specialized training.
In the end, it all comes down to your company’s particular needs. We at craftworks focus on providing tailor-made AI solutions, adapted to your requirements and industry, and then enabling you to continue working with it yourself. We are proud to see the results of our contribution: a rise in resource efficiency, which, in turn, has a huge impact on the overall sustainability of an organization.
Operational efficiency and environmental awareness are the key issues of today and tomorrow, which companies and policymakers must address. What numerous studies and use-cases from industrial practice show, is that artificial intelligence brings great contributions to a resource-efficient business environment.
For that, a certain level of digitalization needs to be achieved. This process can be accelerated by making use of external resources. In the near future, the confidence and acceptance of artificial intelligence will increase (Deloitte Consulting GmbH - Analytics & Cognitiv et al., 2021) and the pioneers will benefit from the competitive advantage through the already developed expertise.
We at craftworks work together with companies as well as with research institutes to support the digital transformation and industry 4.0 while staying up-to-date with innovations in software, Big Data, and artificial intelligence. You can find our guideline for a successful industrial AI implementation here.