Nope. The qualitative data from the questionnaire gets fed back into future workout selection via Adaptive Training. AI FTP Detection is based solely on the AI’s view of the quantitative data of your rides.
As a sense check … could you do 2 x 20 minutes at the FTP it’s given you?
I’d add that VO2 Max work may (read definitely should) be less affected by an overstated FTP than Sweetspot or Threshold work.
Secret Sauce KLAXON!
Indeed. The best clues we have are in the support doc, but they leave out plenty of details (as I would expect):
My 2 most recent AIFTP updates have been 295. I am currently doing SSB HV in week 3 and complete Hermans 4.0 rated tempo/sweetspot workout and rated it moderate.
I dual record and my Garmin 1030 provided an updated FTP of 276 whilst Intervals ICU estimates it at 273 after that workout.
Based on previous experience at this time of year I would estimate around 280 so certainly seems to overestimate for me. I increased from 265 in Jan 22 to 295 at season end early September. I’ve yet to experience a drop in AIFTP estimates despite no racing and generally riding lower intensity for the last 3 months.
I doubt it, not the earlier one of 207, and definitely not 214. I guess I could do 200.
Hard to offer much more by way of (unqualified) advice without looking at your rides. Maybe hit up support@trainerroad.com?
Thanks! I think I may just do a test to check. Or correct it to what I think it should be and go from there.
Of course, this opens the “Does [insert type of FTP test] give me a good representation of my actual FTP?” can of worms. 
If it was me, I’d either do one of the @empiricalcycling long-form tests or, as you say, just manually adjust it but verify it with Gray +5 or some other 2 x 20 workout.
Tbh I’m not stressing about it. My big event is in late July, so I have lots of time. I expect AI will hit the right spot eventually. In the meantime, as long as it feels like the right intensity it should be fine. I’m never going to be an Olympian anyway 
Archery! Start now and we might both make the next games!! 
Maybe curling. My hand-eye coordination is terrible.
No, that is not how Garmin estimates ftp on a 530/830/1030/1040. I get ftp estimates on zone2 endurance rides, tempo intervals, etc. Garmin uses HR, heart rate variability (hrv), and power.
When I have chain gangs (typically level2 with threshold bursts) garmin tends calculate a new FTP so it can’t be as simple as 95% of 20min power.
I’ve been getting “good” ftp estimates from just riding/training since buying the 530 around August 2019. Like I said, sometimes an ftp estimate will come after a zone2 endurance ride. Nice complement to my homegrown HR-to-power based ftp estimates that have worked well since early 2017.
Actually my bad. It was Xert (currently on a free trial) which is now providing an FTP of 274 rather than Intervals ICU.
Intervals ICU is saying 238. Didn’t realise there are options with eFTP calculations so will need to look further into that.
I wouldn’t fully trust the Garmin estimates as I generally only see higher numbers when I perform some full on efforts like a hill TT or strava segment in the 6 - 10 min range.
Sorry to call you out but I had a question that nested into here but didn’t warrant a thread.
Do your post-workout surveys impact AiFTP? I am worrying that if this is the case, doing a low-volume plan (when I could probably fit a mid volume in, but have chosen not to which means I have more recovery between each workout so each workout feels proportionally ‘easier’) is artificially giving me ‘ftp’ increases that wouldn’t be sustainable if I increased my volume.
Does Garmin wrist based HR not work for First Beat?
It wouldn’t be accurate imc but I have only got a cheaper Vivioactive 3 (it will record up to 20bpm lower than my Polar strap).
My Instinct 2 is accurate on my wrist, and it does provide HRV data. One of the big appealing factors was to use it as my main HRM, but on reading the articles in the Instinct 2 and Edge 530 Manuals it looks like it expects a HRM and Power Meter paired for cycling performance metrics. It doesn’t specify HRM for other metrics/modes.
Bother.
@dcrainmaker do you know if this is the case? Is a chest HRM required for cycling performance metrics?
My AI FTP is very close to my Ramp test FTP.
Both seem about right for my training sessions, and I very rarely fail a session.
But I think I can’t manage 2x20 minute intervals at my given FTP.
My inability to maintain threshold for long enough is probably why I’ve disappointed in my last 2 races.
I’m tempted to experiment and do the 2x 20s at my given FTP and see what happens, apart from it being a good workout. My current plan doesn’t have much of those longer threshold intervals.