Growing up in a family that owned a small-scale iron and steel fabrication business, Aniruddha and Avra Banerjee were aware of what sudden machine downtime could lead to. “We always followed constrained deadlines but every time a machine broke down and production halted, we couldn’t deliver on time,” says Aniruddha.
While trying to fix this problem for their family, Aniruddha, a software engineer, and Avra, a hardware expert, designed a product to analyse and predict equipment inefficiency in real-time. After successful trials, they launched Switchon, an AI-enabled industrial IoT platform, so that other small manufacturers could benefit as well.
The start-up offers software-powered sensors which can be attached to the machines in an assembly chain to monitor temperature, energy and vibrations. These are complemented by software, which uses artificial intelligence to create algorithmic replicas of the machines and their components. While the sensors track machine performance, the software helps to identify and, hence, avoid or resolve technical glitches.
The Bengaluru-based start-up, which was incubated by Nasscom Centre of Excellence for IoT and Axilor Ventures, proved its worth to its first client, a major biopharmaceutical company, while working on its chiller plant (used for cooling drugs). That helped it secure its next client, Stanzen, an auto-ancillary that supplies automobile parts to companies such as Toyota, Hyundai and Maruti. “Ancillaries usually use a 500-tonne press, which is critical to their assembly line. Any breakdown during production translates into a delay of atleast five days,” Aniruddha adds, pointing out the difficulties faced by manufacturing units.
Switchon claims 85-90% accuracy in predicting machine failure and is striving to improve it further. Its sensors are priced around Rs.30,000 per node and the software subscription (available in one, three and five-year licences) can go up to Rs.120,000 per year.
The start-up also wants to transform itself from being just a provider of sensor devices to one that provides a complete range of services to address unpredictable machine downtime.