If you've uploaded a ride on the paid version of Strava in recent days, you probably noticed a new feature: Athlete Intelligence. Following the introduction of Dark Mode and Quick Edit, this new addition stands out, especially for the often hilarious texts it generates. This immediately raises the question: what exactly is Athlete Intelligence, and what practical uses does it have beyond just providing a good laugh?
Athlete Intelligence is an AI-powered feature currently available as a public beta for Strava's paying subscribers. Its goal is to help you better understand and improve your performance by converting complex workout data into simple, actionable insights. The feature analyzes your activity from the past 30 days, providing detailed insights into metrics such as pace, heart rate, elevation gain, power output, and Relative Effort—a Strava-developed measure of workout intensity.
So much for Strava's own explanation. What the user now sees in their activity is a text generated by Strava that, at first glance, appears to be a mix of some data points and phrases drawn from the description you yourself added to your activity.
It's still too early to fully assess the feature's value to users. The initial impression, however, is that Strava does manage to identify trends and milestones in performance, potentially highlighting valuable insights that you might otherwise overlook. For instance, spotting an unusual heart rate or power output could be beneficial. Not everyone tracks their progress in apps like TrainingPeaks or Garmin Connect, so if you primarily use Strava and aim to train more effectively, these analyses could be useful. The question remains, though, whether users appreciate receiving data in the form of AI-generated text. A few straightforward graphs or tables would likely be much clearer.
Another challenge is that Strava doesn't know your intentions. If you've been enjoying relaxed, short rides for a month, you would naturally expect a trend change with your first endurance ride or intense training session. Moreover, the averages Strava uses for analysis aren't necessarily helpful. It's well-known that average speed can be misleading, and average power isn't much better unless it's considered in the context of the activity type. Pedaling at an average of 200 watts during interval training is quite different from doing so during a time trial or endurance ride.
Let’s not dismiss Athlete Intelligence too quickly; there's hope that Strava will gather substantial user feedback and refine the feature. At present, it doesn't seem like a strong reason to become a paying subscriber. Personally, I’ve just begun a new training cycle and am eager to see what insights Strava will offer in the coming months and how they will compare to the other apps I use. If I'm not convinced by then, Strava provides the simple option to opt out of this feature.