The company specializes in state-of-the-art visualization solutions having 3600 employees and being present in over 90 countries.
In pursuit of continuous customer satisfaction, the company's devices are cloud-connected to enable data to be gathered on product usage, component status, settings and presets, and environmental parameters. The company had an analytics pipeline in place to process this data but wanted to explore further possibilities with Google Cloud, machine learning and open source software in general.
By transferring the current analytics pipeline to python, extending it with deep learning and building in a self-learning module, the lifetime of certain components in the company's products can be predicted for more optimal maintenance scheduling and prevention of system downtime.
By creating a ‘digital twin’ of the company's products, the individual usage and environmental contexts of each asset can be taken into consideration for more accurate predictions of wear-and-tear and maintenance requirements, rather than routine maintenance scheduling and periodic replacement of parts. These digital twins are also ‘self-learning’, meaning that they continuously validate their predictions against reality and embed that feedback in their future predictions for continuous improvement. By learning the specific environmental context and usage of assets, alternate product recommendations can also be made, where appropriate. The benefit of this is lower maintenance cost for the company, minimized downtime for their customers and a more efficient service all-round.