How AI can help you get into the best shape of your life!

I have less than a month before my trips to London, Hawaii and China! With that in mind, I’ve been searching for the right diet and workout plan. I don’t have a personal trainer, but thought I couldn’t go wrong with the 2-pound rule from this LiveStrong article. To lose 2 pounds a week, I would cut out 1000 calories per day 1. Assuming I stick to a 1200-calorie diet (which has been a challenge because I LOVE FOOD) and normally burn 2000 calories per day without doing much, all I’d need to do is burn an additional 200 calories each day through exercise. Seems simple enough, but it’s been 2 weeks of on-and-off barre, yoga, spinning and 30-min treadmill sessions. And I still weigh the same! Of course, body weight is just one measure and doesn’t account for progress in the form of reduction in body-fat percentage and overall fitness level. However, what I’ve noticed is, I’ve had to take a couple of breaks. I didn’t work out much before this summer and experienced a lot of delayed onset muscle soreness (DOMS). The group classes I took had good instruction and helped me stay motivated, but they were not tailored for someone who needs to step it up gradually. I know I’m not alone in my desire for individualized workout plans. Research shows that depending on one’s fitness level, body composition and medical restraints, there are specific routines that would be more effective in reaching his or her goals. Nevertheless, the cost of a personal trainer is still very high. Hence, that’s why an AI-driven solution that provides customized workout plans is useful for anyone interested in becoming fit.

The Little Big Idea:

Smart Scale and integrated web/mobile application that: (a) tracks users’ progress against their health and fitness goals and (b) leverages their workout history and research from physiotherapists and sports scientists to curate exercise routines that are personalized and highly effective. It does this through an algorithm that predicts, day by day, the type, intensity and duration of physical activity that would help users reach their desired outcome. These predictions will be based on both users’ live performance metrics (e.g. weekly BMI & weight loss, gains in flexibility, muscular endurance, etc.) and insights generated from crowdsourced data collected through the app. Essentially, the more people who train with the app, the more it actively learns from users’ experience, what works and what doesn’t. The application should do the following:

  • Perform an initial physical assessment on users and recommending a starting point. New users are asked to fill out a questionnaire about their medical history, diet, hobbies, barriers to exercise and goals 2. In addition, a series of tests that measure endurance, strength, flexibility and speed will be conducted to assess the user’s fitness level 4. So, whether you’re ready to do squats with 125 pounds or something more low-impact, like a 20-min elliptical interval workout, the app will figure it out!
  • Provide workout videos that you can try out either at home or at the gym. This will be dependent upon your preferences and information from the initial assessment.
  • Introduce additional variety and challenge to help users reinforce strengths, work on deficiencies and avoid injury 3.
  • Adjust workouts to take into account schedule changes as you go along.
  • Correct a user’s positioning. For staple workout exercises, such as squats and planks, the app will provide live feedback to ensure that your form is correct. Through the camera, it compares where your joints, arms and legs are, along with the angles they create, to the “perfect” form 5.

Are there similar ideas out there for you to check out?

  • Freeletics is a fitness app that assigns a personal training plan to each user. It recently integrated AI into its Bodyweight Coach feature, which now provides interval workouts that tackle specific areas of the body more effectively 6, 7, 8.
  • Sworkit guides its users of all levels – from beginner to advanced – through video workouts that can be done at home in as little as 5-15 minutes. The user can customize these workouts to focus on specific areas of the body, workout type, difficulty, etc. In addition, a real, personal trainer is also available to answer questions and help you choose the right workout 9.
  • Perfect Squat Challenge (by Kaia Health) uses AI-powered motion tracking and correction technology to help users perfect their squat. It does this by gauging the relative positions of limbs and joints and the angles between them, comparing them with what physiotherapists would describe as the “ideal” squat. While the technology is very impressive, the app is limited to squats, but has the potential to help users attune to more exercises 10.

So what do you guys think? Any recommendations and suggestions for a workout plan are also welcome!












Continue Reading