Skip to main content

Hi! Thanks for including us in this beta!

 

I have a couple of concerns upon first use: 

  • Much of our guidance needs to be 100% accurate and specific, especially in the case of Trust & Safety issues.  ​​​​​I’m nervous that Ai will potentially miss the exact verbiage or steps that need to be followed. Would there eventually be a means of shutting off Answers for certain content? Is there a way to check its accuracy?
  • Our guidance is mostly written out in steps. Ai summaries lose that format, and I’m concerned steps will be missed. 
  • Our cards have guidance for multiple teams, and will specifically outline which team should be using which guidance. With the Ai summary, the “audience” element is lost, which risks a lower-tiered agent using guidance meant for a higher tier or different team. 
  • Sometimes there are a few sources that show up. Is it possible to know where exactly Answers is pulling its summary from? 
  • I love the idea of eventually using Answers to search across synched apps. How will searchers know if content is verified for use if it’s coming from platforms that don’t offer verification? 

 

Thanks again! 

 

Hi Reba! Thanks for the feedback here, super helpful. A few thoughts: 

  • Totally understand why exact verbiage, step-by-step instructions, and audience specifications would be important here. I’m checking with our data science team to see if they have recommendations here. 
  • The Answer you see in a situation where multiple Cards are cited as a source would be pulling information from all of the cited Cards.
  • Similarly to how Cards are cited as sources today, any sources from outside apps will be cited as well! Of course, other sources don’t provide verification, so there will be an added layer of discernment needed on the reader’s side if they see a non-Guru source cited. 

Reply


Terms & ConditionsCookie settings