The benefits of artificial intelligence (AI) in daily life and across many industries are immense, and the ski industry is no exception. Ski resorts are adapting to this technology, improving guests’ experience and advancing safety.
PyxisAI (pronounced pix-ess) is one such technology. Formed to bring cutting-edge vision and improve ski lift operations, this video-based technology monitors and detects abnormal behavior when a guest is loading onto a lift or getting off one.
Scott Queen, founder and CEO of PyxisAI, developed the product after seeing a little girl on a ski lift ahead of him a few years ago. The little girl failed to unload from the lift at the top. The lift attendant was distracted and did not see the child until Queen yelled to get their attention. The child had jumped off the chair as it rounded the bull wheel on its way to the bottom. At this moment, he was convinced that AI technology could have watched over the skier and prevented this potentially dangerous situation.
“We purposely started small and only had our system at Winter Park Resort in Colorado because we wanted to be able to provide them with ‘white glove’ service. We had a very successful season, improved the system greatly (both from a software and hardware standpoint), and addressed many real-world use cases to improve and harden our technology.”
Rachel Lee, PyxisAI
Queen built and trained his AI model with basic data to test whether his hypothesis would work. When it did, he filed for patents in 2019 and launched PyxisAI, LCC in Fort Collins, Colo., with Sam Shapiro, the chief operating officer and general counsel, and Rachel Lee, who serves as chief product officer for the company.
“Our solution employs computer vision which is that branch of AI that allows computers to see,” said Shapiro. “That is, to act like a virtual lift attendant/operator that is always on, always alert and never gets distracted or tired. Specifically, the PyxisAI system knows when the lift needs to be slowed or stopped to assist a ski resort’s guest to properly load and unload the ski lift.”
How it works
Named after a constellation in the southern sky representing a mariner’s compass guiding navigators and sailors, the PyxisAI system can be configured to alert resort staff by issuing a visual alert, an audible alert or both. It can also be set up to slow or stop the lift in unsafe circumstances.
Essentially, it works like a hyper-vigilant lift attendant who watches guests load and unload ski lifts, and alerts staff when a guest is at risk of failing to load or unload the ski lift properly. Using AI and computer vision, this patented technology is designed to provide a backstop for human lift attendants.
“Importantly, the system cannot speed up the lift or restart a stopped lift, which eliminates the risk that the system would make the lift less safe,” said Shapiro.
“It also can be configured to be more or less reactive, depending on the needs of the ski resort at any particular lift. For example, if the lift has more beginner skiers and/or more children riding the lift, we can make our system more sensitive.”
The use of the system benefits ski resort operations in many ways. It is a backstop for staff who may not see an incident developing. It can also free up extra staff, who would otherwise be needed at a lift, for other higher-value guest services at the resort.
“It can help alleviate the stress placed on resort operations due to lack of staffing and allow lifts to open that otherwise may be closed by using our system’s extra set of eyes,” said Queen.
He says that attendants and operators are often asked to do many different things in seconds as lifts are loaded and unloaded.
“Human error is inevitable. It’s impossible – even for the most attentive employee – to see and hear everything that is going on during the dynamic environment of lift operations. When lifts are not stopped or slowed in a timely manner, people can get hurt, sometimes catastrophically. Not all lift incidents can be prevented, but the PyxisAI system is designed to mitigate those that could be.”
Unnecessary lift stops
Stopping lifts unnecessarily or stopping for long periods of time are pain points for ski resorts. Throughput slows when the lift stops and wait times grow almost exponentially.
“No guest likes being on a lift when it stops and many have a fear of lift stoppages,” said Queen. “Naturally, guests want to spend their time skiing and riding, not stuck in line or on a lift. Superfluous stopping of lifts affects the overall guest experience and guests express this displeasure with their wallets. They will go to ski areas where these frustrations are less likely to occur.”
“Specifically, the PyxisAI system knows when the lift needs to be slowed or stopped to assist a ski resort’s guests to properly load and unload the ski lift.”
Sam Shapiro, PyxisAI
PyxisAI can help this in two ways. For example:
- More timely stops when necessary: When lifts are not slowed or stopped promptly, the ensuing incident typically results in lift stoppages that are longer than otherwise would have been needed.
- Institutional memory: Another source of unnecessary lift stops is staff overreacting and hitting the stop button. New staff tend to be conservative, stopping lifts frequently if they aren’t sure things are correct. With time and training, they learn when stops are truly needed.
“PyxisAI, however, does not have to learn every year,” said Queen. “In fact, it gets better with each year. A resort that uses PyxisAI can train their new staff to rely on PyxisAI and learn from it regarding the need for lift slows and stoppages. This results in fewer stoppages, especially with newer staff.”
Queen, Shapiro and Lee worked tirelessly to have the PyxisAI system ready for the 2021-22 ski season, its first season with a commercial operation.
“We purposely started small and only had our system at Winter Park Resort in Colorado because we wanted to be able to provide them with ‘white glove’ service,” said Lee. “We had a very successful season, improved the system greatly (both from a software and hardware standpoint), and addressed many real-world use cases to improve and harden our technology.”
Industry support
Queen, Shapiro and Lee are delighted that the industry has been overwhelmingly supportive of the PyxisAI technology and efforts to improve lift operations.
“Shortly after we first launched in 2020, the pandemic shut the ski industry down, which we feared would halt our ability to gather video in the field to train our AI models,” said Lee.
“However, numerous ski areas from around the U.S. shared videos that they had archived to allow us to train and improve our AI models, even though the ski areas were closed. This allowed us to continue full speed ahead!”
“PyxisAI, however, does not have to learn every year. In fact, it gets better with each year.”
“PyxisAI, however, does not have to learn every year. In fact, it gets better with each year.”
The National Ski Areas Association has also been an invaluable supporter, inviting the trio to speak at various conferences and events, and publishing an article highlighting the technology in 2020.
“We would also like to give a huge thanks to Winter Park Resort and Alterra Mountain Company, which have seen the value in what we have been developing from the beginning and have allowed us to gather video data, as well as test and then implement our system in the field,” said Shapiro.
As for the plan for the next several years, Queen says with AI, the goal is always to make it smarter.
“We do this by working on the AI model architecture and by training our models with more data. This will be one of our main focuses moving forward. We also plan to continue to expand our patent portfolio internationally and hope to work with some strategic partners to allow us to scale more rapidly worldwide,” he said.
