IBM Watson, the big data machine, is following the money and getting into big industries such as healthcare and finance. The uses of a supercomputer in these areas makes sense. It is far better than humans at crunching numbers fast, and can instantly run detailed algorithms to provide insights hidden in large data sets.
In recent years another area we have seen a proliferation of data is personal health and fitness. For the endurance athlete, surely the training data that has been collected can be used for more effective training? The current solution is to get a personal coach to analyze your past training to inform future sessions. Coaches are great, especially for discussing technique. However, they do have some limitations. Firstly, they are expensive, especially if dedication is desired. And secondly, while many are data savvy, few can undertake detailed analysis of personal data and apply the latest sports science to tailor training plans.
We now see the emergence of online coaching sites, where software such as TrainingPeaks allows coaches to send training plans to athletes, track how they perform, and provide feedback / amend training plans. Despite these advancements athletes are still looking at $100-150 per monthly training block or $150-300 for a 12-week Ironman training plan, that is generic, and not tailored to them. There must be something better…and there might just be.
Everything relies on the underlying ability of AI to analyze big sets of data, and ‘learn’ as more specific data is added
Meet the artificial coach, ready to push you to your limits based on the latest sports science. What is such a coach and how can it improve performance? As the name suggests, rather than a human, a computer structures and tracks performance. Everything relies on the underlying ability of AI to analyze big sets of data, and ‘learn’ as more specific data is added. Here are three of the biggest potential benefits.
- A truly personalized training plan without the need for a lab. There are so many variables to optimize for when creating a plan – an athletes race goals, weekly training availability, weather considerations. A computer can do this at a touch of a button. The real power however, comes from overlaying personal fitness data. Rather than have a human make broad judgments about the type of athlete you are, AI will use large data sets to pin point what works for you. Enough personal data analyzed against a large enough comparator set should yield very good results for key metrics such as VO2 Max, lactate threshold, functional threshold power, and heart rate training zones . These will then form the basis of a training plan and can be dynamically updated as training proceeds and the athlete gets fitter.
- Post analytics and real-time corrective action. After every session, AI can provide you feedback on how you did against the plan and where you could improve. Think Strava on steroids with much more in-depth analysis and the ability to automatically update your training plan. Even better, real-time feedback will help you get the most out of every training session. All the data your garmin is collecting will be instantly assessed and your headphones will be telling you to shorten your stride, or give you quicker run splits if your numbers are good. The potential training benefits of this real-time adjustment versus sticking to static plans are huge.
- Dedication to you (and your wallet). All these services are delivered direct to you on your smart device. No need to wait for your coach to provide feedback days later when you’ve moved on to your next training block. The ‘costs’ of a coach (travel time, need to align calendars, the wad of cash you hand over) are no longer needed. Therefore, the athlete gets a more robust, faster service at a lower cost.
For the moment coaches can breathe easy, the data analysis and slick user tools are not in place to totally disrupt their value. For those able to tap into the online market they can create scale for themselves.
But big data and AI is coming to personal training. It might not be IBM Watson, but for someone there is a valuable enough market out there. To give impetus a new tech-savy player maybe needed to take out cost, offer revolutionary customer experience and get a big piece of the personal training market. Going forward, it may not be the ex-professional who can give the best training advice, but the data scientist. We look forward to seeing what the data geeks provide us!