I did Antelope -3 outside today. My last set ended up with me a 20min ride from home, and I opted to stop at a cafe for some breakfast on the way back.
Historically I’d leave my garmin running while stationary. This feels most “honest”.
But does that play well with the AI algorithm?
Or should I pause the recording, and resume when I set off again (this is what I chose today).
Or actually stop and save and then have a separate short ride afterwards to upload?
Or does it just make no odds at all 
IDK what’s best wrt AI, but for a long-ish stop like that, I do the last option, stop and save the workout, then record a separate ride afterward.
If your cafe stop was after you finished your TR workout, I’d just leave your Garmin as you normally would.
This is, of course, assuming that your cafe stop wasn’t wildly long.
If you have any issues with how TR AI assesses your rides, just reach out. I’d rather tweak the product if we have to than make all of our athletes do something that doesn’t make sense out on the road each day. 