Feedback is central to products that convert. A strong feedback loop, sprinkled in with discipline and excitement will drive you to a successful product.
I used to watch recordings that my team would send me, delivering valuable feedback that would help me know where to go next.
Nowadays, I drop the video into Gemini and ask it for a list of feedback points that I can simply copy and paste directly into a spreadsheet.
It saves me:
10+ minutes of watching the video
5 minutes writing down each point
5 minutes translating it in a way that makes sense
But, Gemini can do all of this for me, much faster than I could ever.
Now, I’m asking myself, why not just remove the copy/paste part- isn’t that even easier to automate? It might take me 5-10 minutes of copy and pasting, but couldn’t I use AI to build something in 15 minutes that does this for me?
BUT, the bigger question is, what am I losing now that I’m automating this process?
Does AI capture the big picture?
Are there any human emotional aspects not being seen that should be?
Does listening to the words, writing it down, cement what’s happening?
Does this de-humanize the feedback?
I’d be curious to see if the positives (saving time 99% of the time) outweigh the negatives, especially as we automate more and more.
The challenge with feedback is separating what’s said from what’s felt.
The rational parts point in a direction and can be easier to translate into engineering specs. The emotional parts reveal what the interface is missing. That’s where the signals create the most value, but are often not direct.
In day-to-day work, this is what is called cognitive empathy:
Finding a different path with feedback can be tricky… but it’s this friction that can provide the most value. Understanding why something should be different is a trust building execise, and requires building influence.
Good questions about the drawbacks @ben. I think it depends on the context whether you’re worried about the output AI delivers: recordings like interviews and moderated usability studies do rely on picking up on the intricacies of facial reactions and emotions, which we know that AI is not as proficient with.
However, recordings like you’re referring to here, where grounded feedback is being provided about next steps for a product, I think the AI is super useful (and trustworthy) for recounting details and discussions. I’ve found several times that it helps pick up on small comments or specific examples provided that can get lost to memory or handwritten notes.
