Based in Kansas City, MyAnIML is developing an artificial intelligence-driven solution for early disease prediction in animals, starting with cows. The company recently spoke to IHS Markit Animal Health deputy editor Sian Lazell at the Kansas City Animal Health Corridor Digital Animal Health Summit about its future plans.
MyAnIML’s platform is comprised of camera technology and a cloud-based system that analyzes images of a cow’s face and muzzle.
Chief executive and founder of the firm Shekhar Gupta explained: “The way the process works is we take a picture of the cow’s face and extract the muzzle, then we compare that to a previous image and if there are any changes. The system then pinpoints what disease or illness that change belongs to, and notifies farmers if a cow is at risk of becoming sick and with what disease.
“These muzzle changes can be picked up before disease impacts the body, before you can see a cow is sick. That’s important because by the time you see a cow is sick, symptoms of the disease are already affecting it. At that time, you’re just detecting disease and letting somebody know, versus actually doing prediction.”
In July this year, the company worked with veterinarians to prove the technology worked in prediction of pinkeye in cattle, which is also known as infectious bovine keratoconjunctivitis. Mr Gupta highlighted the disease is highly infectious, with 90% of cows in a herd likely to be infected if just one cow is sick.
The company currently has around 1,500 cows in its system and has shown an 88% accuracy rate in distinguishing sick versus healthy animals. As MyAnIML continues to collect images and build its dataset, the company’s goal is to boost the accuracy rate to the low to mid-nineties. Mr Gupta told IHS Markit Animal Health the algorithm the business has designed suggests around 2,500 animals would be needed in a dataset to reach its improved accuracy goals.
Mr Gupta said: “Our focus is on the face and muzzle – it’s a very open area and there are no other companies with patents in that area at present. We know what a healthy muzzle looks like, we know what the muzzle of a sick cow looks like, and we also know the time difference between a muzzle that is starting to change and one that is fully changed. We have about two days lead time to tell an owner if an animal is going to get sick, so they can isolate that cow. That will not only save them money in treatment costs, but also their revenue will go up.
“Outbreaks happen because before you can visibly see an animal is sick, it has already infected other animals around it. Those animals are then sold to other farms or ranches where they infect more animals. So, we’re really looking at the predictive market, where we can give about two days – if not more – advanced notification.”
MyAnIML started a trial with West Texas A&M University in September, which is expected to run until April/May 2022. The company is looking to add other diseases to its dataset, such as bovine viral diarrhea, infectious bovine rhinotracheitis, pneumonia and other bovine respiratory disease complexes, influenza, shipping fever and mastitis. The university has around 4,000 beef cattle and 500 dairy cows.
Mr Gupta remarked: “At the end of the trial, we’ll have a very good understanding how each disease changes the muzzle and we’ll have a very good dataset to reflect on.”
MyAnIML’s technology is not licensed from or to any other entities – it has been developed and is currently fully owned by the company. Mr Gupta himself comes from a “very highly technical” background, having previously held leadership roles at various businesses in e-commerce, telecommunications, monitoring software and other areas of technology. He is also currently involved in blockchain consulting for the Kansas City area, while acting as chief technology officer at both Assun (data analytics, machine learning and blockchain products) and Interacshn (machine learning and artificial intelligence software).
The MyAnIML team is also backed by direct agriculture experience, with a sales leader that is a fifth-generation rancher and a business development leader that is also a fifth-generation farmer. The company also benefits from two advisors that are cattle owners themselves.
Other species, potential exit
Further down the line, MyAnIML has plans to expand its technology in other species and is looking at securing an exit for the business.
Mr Gupta said: “There have been people in the swine industry, as well as chicken and turkey farmers, that have reached out and asked me if this technology can be used in their animals. Especially in the swine industry it becomes really hard, because you’re taking a picture of a pig and it is hard to grab the image of their face and muzzle in a distinguishable way. The way their snout sits out means their face and muzzle kind of hides behind it. Those are some really challenging things that would make it hard to take a good picture.
“One of my longer-term plans over three to four years is to get into the companion animal industry, including equine. I have no desire to go into the IPO realm, so will be looking at exiting after that. An offering is a very long-term process and it’s not something we’re looking to do.”
MyAnIML has been involved in the Kansas City Techstars program, which Mr Gupta described as one of the largest accelerators in the world. Along with private backing from the co-founder himself, Techstars has enabled the firm to get to its current stage. The company is now undertaking a funding round with the aim of raising around $750,000.