During the past year or so I have tried out Garmin Connect+ and a paid Strava subscribtion. I was hoping to get some decent extra's and ai feedback but it was quite disappointing.
Most of the ai conclusions both in Connect+ and Strava where poor or even entirely off. It's very disappointing that for example intervals or sprints don't get recognized withour manually pressing lap buttons or whatever. Current state of the art ai should be able to get this out of the raw data my watch produces.
Today I exported some fit files and after converting them to CSV imported them in the free version of Google Gemini. I have been using Gemini for some time to get better workout advise and it works quite good. Today I asked Gemini to analyse two runs. A hill sprint session yesterday and a half marathon slow long run today. I asked Gemini to give me it's best shot because I am considering moving to a paid subscription. The results where quite good and what's even better you can add conversation to improve the analysis on the fly.
Now why am I posting this here?
First of all I think it could be helpful to other runners who like me love deepdiving into their data. Either for fun or better results.
Secondly I would like to give advise to Garmin product management to really accelerate on the ai path. Why not build a plug-inn for ai models so that it will be easy for Garmin users to work with their data in other ai tools. There lies a chance for Garmin to beat competition in ai integration or on the opposite if you don't Garmin might loose it's leaders position in the near future.