Less Mucking Around

In her early 20s, Sarah Parsons worked as a groom at a racetrack and her job, she explains, "was behind the pitchfork." She started work at 5 am and by 10 am would be finished "feeding the horses, cleaning the stalls, bathing the horses, grooming the horses."

Now, at 60, she hires people to do the work. Or tries to, anyway.

"It's a very difficult job," she says. "It's manual labor. You have to do this all the time, whether it's raining or snowing or 100 degrees out here. It's very hard to find people to do this. You usually end up with very young girls who are madly in love with horses."

Kassidy Edwards, 13, cleans up after Rosa Del Paraiso, whom she was preparing to ride during dressage events last week at Santa Fe HIPICO. Local AI specialist Sabri Sansoy's newest venture, Muckerbot, would lend robotic skill for larger-scale manure management. (Julia Goldberg)

Kassidy Edwards, 13, fits the bill. On a very hot Friday, she and her sister Ellie, 15, were cleaning up after Rosa Del Paraiso, Parsons’ horse, at equestrian center HIPICO Santa Fe. They had driven in from Lubbock, Texas to participate in Dressage at Santa Fe I & II, a three-day event from the Santa Fe Dressage Association.

The Edwards sisters, indeed, said they did not mind cleaning up after Rosa Del Paraiso (the horse, also, did not seem to mind).

Cleaning up after one horse is one thing. Finding people to clean up after hundreds of horses is a horse of a different color.

HIPICO co-owner Phyllis Gonzales says even before the COVID-19 pandemic, finding help was an "ongoing struggle" and it's grown harder since. An experienced "stripper," she says, can clean up a stall in 15 or 20 minutes. "And that's a good one, somebody who knows what they’re doing and has been doing it a really long time." Sometimes, she says, HIPICO will need as many as 200 to 300 stalls turned over in 48 hours and "we’ve struggled mightily to find [people] to do this."

Enter Sabri Sansoy, a Santa Fe roboticist and Artificial Intelligence specialist. I first interviewed Sansoy about three years ago, pre-pandemic, when he had just moved home to New Mexico and had started an AI meet-up group. His career has included a wide variety of robot-related work across sectors, including film and agriculture. For example, he built sentry paintball gun robots that used deep learning to recognize human targets and fire paintballs at them for Ridley Scott's Scott Free Productions. On the other end of the spectrum, he's worked on projects to incorporate robotics and deep learning for manual labor like picking oranges.

His interest in the latter sector prompted his new foray: robots to help people clean up horse manure. The idea grew out of a conversation Sansoy had with Santa Fe veterinarian Doug Thal, a childhood friend, who raised the issue of "manure management," when Sansoy asked him how AI could be of help in the veterinary field.

From there, Sansoy teamed up with former Deutsch LA colleague Fred Leveau, who works in product and design, and they began planning Muckerbot.

The two participated in Santa Fe Innovate's business accelerator program last year, placing second in the final judging round. The program took place virtually, allowing Leveau, who lives in Atlanta, to participate. Santa Fe Innovates Founder Jon Mertz says the virtual format created diversity in both participants—attracting people from all over the state and beyond—and ideas. The pandemic's impact on the employment landscape has definitely infiltrated the entrepreneurial sector.

"I think that's the great thing about entrepreneurs," Mertz says: "They’re flexible. They can see some of these trends unfolding and they can jump in with an idea and see if it has traction or not."

In the case of Muckerbot, he says, "it's very innovative. It [was] something I hadn't thought of, but they definitely have the experience in that marketplace. And I know when they talked to potential customers that they were able to iterate and get some good insights."

Sansoy and Leveau surveyed 50 operators of different types of commercial horse operations, visiting five in New Mexico—HIPICO, among them. Of the four opportunities in manure management—collection, disposal, storage and utilization—they decided the first, collection, required the most labor, had thus far been the least impacted by technology and was the problem they were most equipped to solve.

According to the most recent estimates from the American Horse Council from 2017, there are approximately 7.2 million horses in the US. They each create around 50 pounds of manure a day. That adds up to more than 3.5 billion pounds of manure each day, with a declining labor force to clean it up.

While there are some automated solutions for other animals (cattle, for instance, and cats), currently, Leveau says, "there's nothing in the autonomous solution for horses." Sansoy notes that in a visit to Dubai he also learned the camel industry has a similar need.

While the demand exists, Sansoy and Leveau say some technological challenges remain. Machine learning is up to the task of detecting items that need cleaning and removal, but in this case will need to do so in a dirty environment among easily spooked animals.

"How do you make the system quiet? Do you do this when the horses aren't in the stall?" Sansoy says. "But sometimes you have to do while it they’re in the stall…so there's lots of challenges."

While Muckerbot would help with the dirty work, Sansoy and Leveau say it will be more of a collaborative robot—cobot—and still require human labor, but of a less back-breaking variety. Cleaning out stalls is just one facet of a larger job and no one thinks anyone working in large-scale stall clean-up would miss the job.

"The way we want to position this is that this robot is going to be a multiplier of the labor force in that environment," Leveau says. "Unless you have one horse and you see the task of scooping poop and being there as a connection to your horse…and we do see when people have one or two horses [that] can serve as a connection. When you have more than eight…it's just get in and get out…ideally, that's the promise of technology—we’re removing the terrible backbreaking activities and making life easier for you."

Indeed, Gonzales says, anything that makes the job more efficient for the people they hire to strip, feed and "do all the things that help the riders get themselves together for the show ring" would be welcome.

"We wish him good luck," Gonzales says of Sansoy. "We love it when he comes out here to take pictures of manure."