Fastems Utilizing AI on Unstructured Data to Gain Smart Insight for Automation Solution Usage
What was the challenge?
Fastems is a world-leading supplier of CNC automation solutions with over 40 years and 4500 installations of experience with over 100 machine tool and other device brands. They solve the fundamental challenges and inefficiencies around variable batch (high-mix) production, experienced by all industrial manufacturers from local machine shops to global enterprises.
It was realized that the machines that the CNC automation solutions were deployed on have an untapped goldmine – this goldmine is unstructured log data.
We started with a pre-study of a snapshot of log data from some chosen end customers. Before diving into the data we formed some hypothesis on what kind of questions we could potentially answer. The main themes revolved around using AI for predictive maintenance and error sequence analysis, forming a customer index based on machine utilization and other derived metrics, and structuring the data for the maintenance team to have a more systematic approach to root cause analysis.
We were able to validate these ideas based on the data and expert interviews in the first two-week timespan. The result was a comprehensive report on the feasibility of use-cases and suggestions for the next steps towards production deployment and use.
After the pre-study phase we started productionising and further refining the prioritized use-case. This led to the development of a data intensive tool for the maintenance team to be used in their day-to-day work.
Codento has a long history of agile software development, and our vision of the required journey was considered credible. Our proposed approach, from a short pre-study to prioritized use-cases, was a low risk opportunity for the client since neither party knew what could be found from the data, and what AI use-cases can be implemented based on the data.
What were the results?
So far, we have developed a log analysis process to bring structure into the unstructured log files and show graphs and KPIs specified by the maintenance team. This allows for data based analysis of the customer systems and their state. We also started work on the customer index and AI for predictive maintenance use-cases.
In customer’s own words
“In the past we have tried to make use of unstructured data based on our direct needs and requirements but did not really succeed which had made us doubt if we could utilize the data we have at hand at all. This time we wanted a partner to look at the data and show us what could be built based on it. Data analysis and visualization by Codento´s specialists really made the difference and convinced us that there is much gain and we should continue to pursue data visualization and AI further on.”
Head of Services Offering and Customer Success
For more information, please contact: