Roundup of AI Notetaker improvements
We initially released our AI Notetaker to Public Beta on September 25th, and since then the team has been hard at work wrapping up the final details for this beta period. Here’s a quick roundup of what’s been happening over the past couple of months!
Performance improvements to transcripts
We’ve optimized the performance of our transcriptions to better handle the cases where participants are speaking simultaneously.

A nice improvement to transcripts during interviews and debriefs that often involve rapid "back-and-forth" between participants.
Report on employees with AI Notetaker enabled
We’ve added a new field to Employees called “Has AI Notetaker Enabled?” for reporting.

This gives you an easy way to see all your employees whose interviews are eligible to use AI Notetaker.
Better troubleshooting when AI Notetaker can’t record
AI Notetaker won’t record meetings in certain cases, such as when the video-conferencing tool doesn’t admit the bot from the waiting room or when participants join too late to let it in.

You’ll now see in-app messages whenever AI Notetaker can’t record for these reasons (or others), giving you clearer context and helping you adjust for future cases.
Candidate consent updates in the candidate feed
We’ve added better visibility into candidate consent changes by surfacing this activity in the feed on a candidate profile.

Whether consent is updated by the candidate or an Ashby user, you’ll now see a timestamp and a reference to the state of the consent for the candidate.
We’ll have more public beta updates between now and the end of the year centered around consent management, reporting, and a number of UX improvements. We’ve been seeing great results from our current customers around time savings, faster feedback submissions, and more consistent / structured interviews.
As a reminder, AI Notetaker is available to purchase across all plans and each plan includes 500 interviews to try AI Notetaker at no cost. You can enable it on your account under opt-in features or learn more about how it works in our knowledge base.