SwitchOn, a vision AI company, has raised $4.2 million in Series A funding, led by a Singapore-based fund, Axilor Ventures, Pi Ventures, and prominent angels such as Anuj Bihani, Laxmi Narayan, and others.
This investment will allow it to continue its rapid growth in India, scale internationally, hire more workers in sales and technology, and invest in research and development.
The company claims to have helped manufacturing industries reduce defects and improve product quality through its AI system, which is being used by some of the leading manufacturers, including Unilever, ITC, SKF, and others, to train deep-learning quality control models within hours.
- Also read: How is ChatGPT affecting IT employees’ jobs?
According to the AI company, its DeepInspect helps manufacturers identify complex surface defects quickly and achieve superhuman accuracy of over 99 per cent. Moreover, the setup uses less than 200 images to train AI models and enables manufacturers to adopt and completely automate their quality inspection and deploy hundreds of SKUs within days, it added.
‘Zero-defect reality’
“We believe that we have stabilised AI to the point where single-day trials and single-week deployments are finally a reality. We are thrilled to have the support of investors who understand the industrial B2B SaaS space as we continue to make zero-defect a reality for the manufacturing industry,” said Aniruddha Banerjee, Co-Founder, SwitchOn.
While manufacturing companies globally employ more than 25 per cent of their workforce on quality inspection alone, SwitchOn said that they still face over 50 customer complaints per plant every year, leading to over $6.5 billion in brand reputation impact, product wastage, and losses.
“The company has scaled and has several Fortune 500 customers globally. It has unlocked the inspection of complex use cases in automotive parts, consumer goods, and pharmaceuticals which shows the robustness and scalability of the platform,” said Shubham Sandeep, Managing Director at Pi Ventures.
- Also read: Maybe AI needs to write its own dictionary