A beloved horse dies in the night, and a startup is born bringing AI and machine learning to animals – GeekWire
Alexa Anthony has been around horses for much of her life. She grew up show jumping and became an NCAA champion on the equestrian team at the University of South Carolina. Her mother owns a boarding and training facility in Redmond, Wash., and her sister also trains horses.
But it was during a brief period when she was not around her beloved horse Magic that the impetus for her new company was born.
Anthony is the CEO of Magic AI, a new Seattle-based startup focused on bringing computer-vision-based machine learning to animals and humans with a focus on nonverbal forms of data and communication.
The company’s first product is called StableGuard, which is billed as the world’s only artificially intelligent, 24-hour monitoring and alert system for horses. It’s a piece of technology Anthony wishes she had on Christmas night in 2012.
Anthony’s 8-year-old horse, Magic, was stricken with colic on that night. The digestive disorder is a leading cause of death in horses, but it can be treated when caught early enough by administering an anti-inflammatory drug.
“If this happens in the middle of the night when no one is watching, they can go all night with this and in the morning often times it’s too late,” Anthony said. “That’s what happened with my horse. I found him at 8 the next morning and he had been toxic for 15 hours so we had to put him down. But if there was something that had notified me as soon as he was showing symptoms of distress I could have gone out there and I could have given him a shot of banamine and we could have prevented it. He probably would still be here today.”
Anthony got her start in tech after injuring her back during her senior year of college, while training to go the Olympics. She said she realized she was going to have to do something else, and she landed at Bellevue, Wash.-based Intellectual Ventures, helping with startups and spinouts at the patent powerhouse.
It was at Intellectual Ventures that she met Jacob Sullivan, an entrepreneur and engineer who spent 10 years at the company, and the idea for StableGuard took root.
“At the time we were discussing technology portfolios — IV has a lot of patents — and we were trying to figure out business models for applications and almost all of them were human,” Sullivan said. “And the topic came up about Magic and what had happened. And immediately we began thinking, ‘Well, there’s got to be non-invasive technology supervision.’”
Anthony left Intellectual Ventures in May and attended a TechCrunch startup weekend for machine learning to test out the idea. They ended up taking second place among 12 teams who pitched and they also attracted a computer vision expert who had previously worked at Amazon and they met the architect who has written much of the AI for their proof of concept.
Magic AI incorporated in June, Sullivan left Intellectual Ventures in August and the rest is startup history as six people have been working around the coffee table in his basement ever since.
“We think that there’s a lot of data that’s locked up in watching humans and animals do different things,” Sullivan said. “Right now we’re hyper-focused on StableGuard — the application of horses — because they are such a valuable asset, there is a lot of money in that market and it’s completely underserved because a lot of existing technology is just hard to use in that environment. So we think our platform leapfrogs many of those issues.”
Via a video camera in a horse’s stall, StableGuard works by learning to recognize various signals put off by the animal related to everything from stress, comfort related to temperature, food and water consumption, waste evacuation, pregnancy and security.
“Right now we have built machine learning — convolutional neural net — that does recognize horse behavior,” Sullivan said. “The system is a little more complicated than putting a camera in the stalls. We do do that, for the live video streams, but then the machine learning on the backend is designed to create notifications if there is abnormal behavior or track that horse over time — just kind of their habits, if everything is normal or not.”
StableGuard users can monitor everything that is going on via a mobile app.
At Potcreek Meadow Farm, the facility owned by her mother, Anthony said StableGuard is installed in four stalls, including that of her current horse, Elektra.
“We did catch one of the horses acting a little bit funny the other night and so we called my mom and it turns out he was kicking bandages on his back legs because he had been competing,” Anthony said. “And she was like, ‘I’m glad you called me because he probably would have been really sore the next day if I hadn’t taken those wraps off.”
Magic AI is proud to bring bleeding-edge technology in the software industry to the equine industry, a typically small and close-knit community rooted in traditional ways. But the company believes horses are just the start, and they have their sights set on many different applications for computer vision and machine learning in the coming years.
Sullivan said they watch Amazon and Google very closely, and while it makes sense for those tech giants to focus on humans, he wouldn’t be surprised if they eventually get into horses and pets.
And along those lines, Anthony and Sullivan chuckle at the prospect of coming up with another “skill.”
“Alexa, check on Elektra.”