The 2026–27 figure skating season arrives with plenty to watch. Established names are returning to international competition, younger skaters are moving through the ranks, and Alexandra Trusova, now competing as Alexandra Ignatova, is back on international ice. But some of the more consequential changes in skating may be happening away from competition.
Victoria Drazdova, a former professional figure skater, researcher and Head Coach at VSA, has been looking closely at one of them: how artificial intelligence could change the way skaters understand their own technique.
Her interest comes from a familiar problem in skating. Athletes repeat jumps constantly, sometimes performing dozens of attempts in a session in search of consistency. Yet more repetitions do not necessarily reveal why an element is failing. A skater may focus on the landing when the problem actually began during takeoff, or try to rotate faster when body position earlier in the jump is the real issue.
“Experienced coaches rarely look at the fall itself as the whole problem,” Drazdova says. “You go back through the jump and look for the point where the technique started to break down.”
That thinking is informing an AI-assisted analysis tool currently being developed at VSA. Skaters would be able to upload videos of their jumps and receive feedback on aspects of execution including takeoff, timing, body position, rotation and landing mechanics.
Drazdova, who has spent more than 25 years in figure skating and more than a decade coaching, has also researched the use of artificial intelligence and wearable feedback technologies in the sport. At VSA, she works within a large international community of skaters and coaches, giving her a practical view of technical problems that recur across ages and ability levels.
What interests her most, however, is not analyzing a single jump. It is what becomes possible when those jumps can be followed over time.
VSA is developing the system so athletes can retain a history of their attempts and compare changes in technique across a week, a month or a longer period. Instead of relying on an athlete’s memory of how a jump felt several weeks earlier—or comparing two videos buried in a phone’s camera roll—a skater and coach could follow how particular technical patterns are changing.
That could make progress easier to evaluate. A correction may work for several sessions and then disappear. An old habit may gradually return. A jump may feel more comfortable even though one part of the technique has remained largely unchanged. A longitudinal record provides another way to distinguish genuine technical progress from a particularly good day on the ice.
“The higher the level, the more important the details become,” Drazdova says. “One good jump doesn’t necessarily tell you very much. I’m much more interested in what is happening across 20, 50 or 100 attempts and whether the athlete is actually changing the pattern.”
For younger skaters, the same information could serve a different purpose. Children develop quickly, and changes in height, strength, coordination and body proportions can affect technique. Being able to look back at how a jump developed may help coaches understand why an element that was once consistent has changed and where adjustments may be needed.
Drazdova also sees a potential benefit in the way athletes manage training volume.
Jump training places repeated demands on the body, particularly when skaters perform the same element over and over. AI analysis cannot prevent injuries, and it should not be presented as a substitute for appropriate coaching, conditioning or medical guidance. But better information could help athletes and coaches recognize when repeated attempts are reproducing the same technical error rather than solving it.
“If you do 30 attempts with the same mistake, the 30th attempt isn't automatically more useful than the fifth,” Drazdova says. “Sometimes the smarter decision is to stop, understand what is happening, make the correction and then try again.”
It is a relatively simple idea, but it challenges a deeply established habit in competitive sport: equating more work with better work.
In that sense, AI may ultimately prove most useful not by asking skaters to do more, but by helping them extract more information from the training they are already doing.
There is also an economic argument behind the project. Detailed biomechanics and motion analysis can involve specialized cameras, sensors, software and professional interpretation. Those systems have important applications in research and high-performance sport, but they are not part of the everyday training environment for most skaters.
Computer vision creates a different possibility. It cannot replicate everything that can be measured in a biomechanics laboratory, and Drazdova is careful about making that distinction. But if useful elements of technical analysis can be derived from ordinary skating footage, a smartphone video could become a much more informative training record than it is today.
That could matter at both ends of the sport.
An advanced skater may use the technology to monitor very small technical changes that are difficult to assess from one session to the next. A developing athlete may use it to understand mistakes earlier, potentially reducing some of the trial and error that accompanies learning complex jumps.
For VSA, which works with skaters across a wide range of levels, the latter is particularly interesting. Drazdova believes younger athletes should not have to reach an elite training environment before they begin learning how to analyze their own technique intelligently.
The aim is not to promise a shortcut to elite skating. Triple and quadruple jumps still require years of physical preparation, technical development and expert coaching. But better feedback earlier in that process could help young athletes use those years more effectively.
The idea also changes the role AI is being asked to play. Rather than attempting to create a digital replacement for a coach, the technology becomes a record, an analytical aid and another source of information for decisions made by athletes and the people training them.
That distinction matters to Drazdova.
“A coach sees much more than mechanics,” she says. “You know the athlete, you understand fear, confidence, fatigue and what is happening in training. AI doesn't replace that relationship. What it can do is give us another layer of information.”
That may be the more realistic direction for AI in figure skating: less autonomous coaching, more intelligent observation.
As the 2026–27 season unfolds, attention will naturally remain on competition results, new programs and the athletes pushing technical difficulty forward. Behind those performances, however, the way skaters train is becoming more measurable.
Figure skating will always require repetition. Drazdova’s argument is simply that the next generation of skaters may know considerably more about what each of those repetitions is actually doing.
And that could make the difference between simply putting in more hours and making those hours count.
This article was written in cooperation with David Rabinowicz