How does the Fellow Meeting Assistant work?
The Fellow Meeting Assistant is a bot which actually joins your video conferencing link to record the meeting.
The audio from the recording is then sent to our transcription provider (Deepgram) which then produces a transcript from the recording.
Fellow then takes this transcript, processes it into a usable format, with timestamps and speaker labels, and produces a summary of the call.
How does the Fellow Meeting Assistant get added to a call? Is it automatic?
At the moment, the Fellow Meeting Assistant can only be added to a call via explicit action (clicking record) by a user of Fellow while in the call. The recording banner will show up 5 minutes prior to the scheduled start of the meeting.
While we do not support automatic scheduling the meeting bot at the moment, we intend to add this functionality in the future.
Is any indication provided to attendees that the meeting is being recorded?
Yes, there are two ways whereby other participants may discern that the meeting is being recorded:
The presence of the meeting bot itself (including from the video output of the bot), and
The bot sends a chat message in the recording service once it joins, indicating who added the bot and that it is recording the call.
If I share my screen during call recording, will the contents of the screen-share be included in the recording?
Yes. Our bot joins the call in the same way that any other participant would. The recording thus includes the video feed from active speakers, as well as any information presented to the screen at the time of the recording.
Can the recording be stopped or paused in the middle of a call?
Yes. The recording may be stopped by removing the bot from the call. This can either be done via the recording banner at the top of the meeting note (by any user from the owner workspace with access to the note for that meeting), or by kicking the bot from the call via the controls provided by the video conferencing provider. Once the bot is removed from the call, the recording ends.
The recording can be restarted recording after it was removed by re-adding the bot to the call (via the recording banner above the note).
Note that the bot will automatically leave the call:
after all other attendees have left,
after 3 hours from the start of the recording
after 5 minutes of being stuck in a waiting room (for platforms where there are waiting rooms)
Is my data encrypted both at rest and in transit?
Yes. All of our underlying providers encrypt data both at rest and in transit. You can read more about data privacy at Fellow here: https://help.fellow.app/en/articles/6210776-data-privacy
What third-party service providers are used to provide the service?
Fellow takes great care in choosing its vendors, and we ensure that the vendors we select are employing the highest security and privacy safeguards to keep your data safe. We do not send any more information to any vendors than minimally necessary in order to provide our services.
The current additional third-party vendors used to provide the Fellow Meeting Assistant service are listed below. As we continue to develop this features, we may change providers as appropriate (any changes will always be reflected on our sub-processors list)
OpenAI, LLC
Provides AI models for text comprehension and generation.
Data sent by Fellow to OpenAI is not used to train or improve any models. Data may be retained by OpenAI for up to 30 days for abuse monitoring purposes.
Hyperdoc (d/b/a Recall AI)
Provides the underlying infrastructure for the bot which actually joins your meetings. The bot records the call, and sends the recording to Fellow after the call is complete.
Recording data is deleted immediately after successful transfer to Fellow, a process which begins when the bot leaves a call.
Deepgram, LLC
Provides the audio transcription services.
Audio data and transcription output is deleted immediately. Note that this is different from Deepgram's off-the-shelf retention policy, and Fellow has employed special provisions to ensure that data is not retained for training or any other purposes by Deepgram.
Amazon Web Services
The primary infrastructure provider for Fellow. Meeting recordings, transcripts, and summarization results are transferred to AWS after processing from the above third party vendors to ensure we can make them available to you at any time.
Do these third-party service providers retain access to our data?
Once the video recording, transcription, and summarization is processed by the above vendors, it is removed from their systems after the requisite length of time has passed (see above). We have reviewed each vendor's security posture and have determined that they meet our rigorous standards for security and privacy practices.
How long will our calls and transcriptions stored?
For the data retention practices by our vendors, please see above.
Transcripts and summaries are stored indefinitely, for as long as the meeting note itself remains on Fellow (that is, until it is deleted). As for the video recordings, while Fellow does not have a data retention limit on stored meeting recordings, we may choose to implement this in the future to limit storage costs.
Can I delete my call recordings, transcript, or summarization data?
Yes. To delete all artifacts relevant to a particular meeting you can delete the meeting note. The system will then purge all artifacts pertaining to the meeting note, including the recording, transcription, and summary.
If an admin requests that recordings be deleted, all recordings, transcriptions, and summaries will be deleted within 30 days.
I noticed the Fellow Meeting Assistant being used in a call I was in, but I am not a member of the Fellow workspace which initiated the recording, or do not have access to the note. Can I delete the recording?
No. You may only delete the a call recording and other artifacts of the Fellow Meeting Assistant if you are a member of the workspace where the recording was initiated, and have edit access to the meeting note for that meeting.
If you were on a meeting but aren't able to access it in Fellow, check this article for more information.