Milk Marvel: SomaDetect

Optical Sensors and AI Revolutionise the Dairy Industry


Better data for farmers, better milk for consumers, and better lives for cows.  This is the rallying cry for Bethany Deshpande and the company she leads, SomaDetect.  

It wasn’t always obvious that this would be her choice of direction. She didn’t grow up on a farm, she wasn’t privy to the dairy lifestyle & she didn’t study agriculture in college.  She does, however, have a knack for entrepreneurship and connecting ideas, and used this to create one of the dairy industry’s most promising technology startups.  Her curiosity and determination means that SomaDetect can support more efficient & cost-effective milking through the use of optical sensors and the power of artificial intelligence, to detect health and reproductive issues in cows. 

The Beginning

What started as a family conversation about a technology breakthrough, developed literally in the family basement, eventually led to a venture-capital-backed business aiming to benefit the dairy industry. 

“My dad is a biophysicist, who has a fundamental belief that there’s a lot more we can do with optical technologies, light scattering technologies, than what exists today. He’s been developing different forms of sensors for many years, in the basement of our family home, and accidentally discovered that he could see different fat content levels and somatic cells in raw milk,” recalls Bethany. That discovery had the father-daughter team questioning how the sensor technology could be practically applied in the dairy industry. 

Bethany formally established SomaDetect in 2016, intending to commercialise this sensor technology. She enlisted the help of experts in AI, deep learning and computer vision, including her co-founder Bharath Sudarsan, and got to work. While none of the founding team members had great knowledge of dairy operations, they were familiar with using massive data sets and artificial intelligence.  They set out on a journey to understand the potential market and develop a product that would fit farmers’ needs.

“We learned very early on that dairy is an incredibly complex system. Over time, we also realised that when we added deep learning and artificial intelligence to the sensor technology, it offered even more possibilities. It’s been a fascinating, exciting journey with lots of learnings,” says Bethany. 

The Mission

SomaDetect’s mission is simple: To provide farmers with the information they need to make the best possible milk. They do that with a series of technologies that deliver real-time insights, enabling farmers to analyse and act. 

Light scattering optical sensors operate on raw milk collected on a milking line. The sensor, used with computer vision and deep neural networks, detects properties such as somatic cell count and fat content.  These provide important indicators of udder health, which enables mastitis infections to be avoided or treated without recourse to antibiotics. It also helps understand the fat content and hence the commercial value of the milk.  The technology can also detect changes in each cow’s reproductive cycle.  Collectively this helps dairy farmers plan, increase milking efficiencies and reduce costs. 

“Getting computer science and engineering talent on the farm to help improve efficiencies will be critical for the future of ag”

“We’re able to collect metrics about milk quality and animal health by examining raw milk as it comes from individual cows. We share that data with farmers, allowing them to have a much richer data set than they’ve ever had before. They’re able to identify trends or changes from the individual cow level to the pen level, to the whole herd, to catch health and reproduction issues a lot earlier. Ultimately that results in healthier and more sustainable dairy operations,” explains Bethany.

The Path to a Product

While SomaDetect’s mission has always been clear, its path to commercial success has not. While the team knew they had an interesting idea, they weren’t sure how exactly it could be applied to the dairy industry. Would the technology be useful? Could farmers use this data in practical ways? How could the technology be improved? These were the questions that needed to be answered for a successful commercial launch. 

“The first year of the business was really an incredible learning journey. Our team visited more than 150 farms and asked a lot of questions. Can we see your milking equipment? Can you tell us what it takes to raise cows and produce our food? Can we come along in that journey? We were waking up unbelievably early to drive to these farms, and we’d be milking alongside farmers. We got to experience what it took to run those operations,” recalls Bethany.

All those hours of on-farm fact-finding resulted in a prototype built with tremendous farmer input. The SomaDetect team also continued to invest time on the commercial and financial needs of the business. They participated in many entrepreneurial events and business accelerators to expand their network and learn from leaders across tech industries. Eventually, they secured funding to support their commercialisation efforts. As of September 2021, the company has more than 20 full-time employees, raised over $10m from investors and started generating its first revenues within the market. 

“I’m really grateful for the growth that we’re having in ag-tech. I’m incredibly grateful for the farmers that get involved at the early stages. The risk they take goes a long way to develop these new technologies. Without early adopters, these advancements wouldn’t be possible,” says Bethany. 

Throwing the net wider

As a visionary creating new technology for the dairy industry, Bethany feels there’s still plenty of work to be done. Feeding a growing population with limited resources presents an opportunity to innovate and creatively approach challenges. 

“I think data is one of the things we’re monitoring that will help us get smarter and hopefully find more efficiencies in food and livestock production. Automating processes will become more relevant as well. We’re starting to see many more engineering and data-focused jobs in agriculture than there used to be. Getting computer science and engineering talent on the farm to help improve efficiencies will be critical for the future of ag.”

Want to comment?

Please enter your comment!
Please enter your name here