About RealMeeting
RealMeeting is Reality Defender’s deepfake detection solution for live meetings. It helps you verify whether people in a Microsoft Teams meeting are authentic by detecting AI-generated video and voice in real time. You’ll see an alert if a participant is likely to be AI-manipulated.
RealMeeting is designed for high-trust meetings, such as job interviews, leadership discussions, and sensitive internal or external conversations. As AI deepfakes increasingly impersonate colleagues, candidates, or partners, RealMeeting provides clear, real-time signals so you can decide whether to continue or end a meeting before sharing confidential information with an unknown attendee.
How to Access RealMeeting
RealMeeting is made available to you by your company’s IT or security team.
An administrator adds you to your organization’s RealMeeting account.
Once access is granted, you’ll receive an email with instructions to get started.
You’ll be prompted to install or enable the RealMeeting app for Microsoft Teams if it isn’t already available.
Aside from installing the plugin, you do not need to configure any settings yourself. If you don’t see RealMeeting available in Microsoft Teams, contact your internal IT or security team for access.
Roles and Permissions
Your access to RealMeeting is determined by the role assigned by your organization's administrator. There are three RealMeeting roles:
Role | Can run the RealMeeting app | View results |
Admin | Yes | All meetings |
Contributor | Yes | Own meetings only |
Viewer | No | All meetings |
NOTE: If you also have the RealScan Contributor role, you will only see your own meetings even if your RealMeeting role is Viewer. Contact your administrator if you need broader access.
How to Use RealMeeting
Enable RealMeeting
It is recommended to enable RealMeeting for any high-trust conversations, such as interviews or financial discussions.
Start a new Microsoft Teams meeting.
2. In the meeting menu bar, click Apps.
3. Select the Reality Defender app.
If you don’t already have it installed, you may be prompted to Add the app first.
4. Click save to add the Reality Defender plugin to the meeting
5. Once activated, a side panel will open for you as the meeting host.
A RealMeeting detector bot will join the meeting and start scanning each participant automatically
NOTE: All participants will see the detector bot join the meeting, similar to an AI notetaker. Only the meeting host can see the side panel and scan participants.
Scan Participants
Once the RealMeeting bot is enabled, scanning begins automatically.
When the RealMeeting bot joins the meeting, it starts scanning participants automatically and sequentially, based on the order in which they joined the meeting.
Participants are scanned one at a time during the meeting.
Each scan captures a short segment of video and audio and analyzes it in real time.
If a participant has both their camera off and microphone muted, the bot will attempt to locate scannable content for approximately 30-60 seconds before moving on to the next participant.
If a new participant joins while RealMeeting is active, they will be scanned automatically.
After all participants have been scanned once, you can manually rescan any individual for additional verification, if needed.
As the meeting host, you may add or remove the bot at any time during the call.
Read Scan Results
Within a few seconds, you’ll see a result in the side panel for each participant scanned:
Authentic – No signs of AI manipulation detected.
Manipulated – Signs of AI-generated video or voice detected.
Each result includes a confidence score to help you understand how strongly manipulation is suspected.
In-Meeting Consent Message
If your administrator has enabled the in-meeting consent message, the RealMeeting app will notify each participant when they join the meeting before any scanning takes place.
Where to find this setting
Organization Owners or RealMeeting Admins can find this setting in the Reality Defender web UI under Settings > RealMeeting. Toggle In-meeting consent message on or off.
What participants see
When a participant joins the meeting, they receive a message in the Teams meeting chat:
"The host has chosen for an automated agent to attend this meeting to scan participants for signs of AI manipulation."
Participants select Yes or No and click Acknowledge.
NOTE: The consent message is sent to all meeting participants, including the host. This is expected behavior.
How consent affects scanning
In your host panel, each participant's scan availability reflects their response:
Participant response | What you see in the host panel |
Yes | Participant is queued for automatic scanning. |
No | "User has not consented to being scanned." |
No response yet | "User has not consented to being scanned." |
NOTE: The consent setting is applied when the meeting starts. If your administrator changes the setting after a meeting has begun, it will not affect that meeting.
Best Practices
Encourage participants to keep their camera on and face visible.
Avoid heavy virtual backgrounds or motion blur when possible
Ensure at least 5-10 seconds of visible face time for video scanning
For audio analysis, ensure continuous speech (3+ seconds works best)
If a Participant Is Flagged as Manipulated
If RealMeeting flags a participant as Manipulated, it means signs of AI-generated video or voice were detected. This does not require you to confront the participant or explain the result.
Step 1: Rescan Once
Run one rescan to confirm the result.
Rescanning can help rule out temporary issues like lighting, camera quality, or network problems.
Step 2: Limit Sensitive Discussion
If the result remains Manipulated, avoid sharing confidential or sensitive information.
You may choose to pause the discussion or move to a non-sensitive topic.
Step 3: End or Pause the Meeting if Needed
If the meeting involves hiring decisions, approvals, or confidential information, it is recommended to end the meeting or remove the manipulated participant.
Use neutral, non-confrontational language.
Suggested talk track: “We’re experiencing a technical issue with this call. We’ll pause here and follow up separately.”
What happens next
You don’t need to take any further action after the meeting. Detection results are automatically available to the appropriate internal teams (such as security or HR), based on your organization’s setup.
Reminders
You’re not expected to explain scan results to other participants.
RealMeeting is designed to support safe, real-time decision-making during meetings.
When in doubt, protecting sensitive information is the right choice.
FAQs
1. Who can use RealMeeting during a meeting?
Only the meeting host can enable RealMeeting and scan participants. Other participants cannot control the app or view results.
2. Do participants need to give consent for us to run the RealMeeting bot?
Participants follow Microsoft Teams’ standard consent flow when joining a meeting. They must accept the platform’s notification before entering, which covers recording or monitoring. RealMeeting doesn’t introduce a separate consent step.
3. Do participants know when RealMeeting is running?
Participants will see the RealMeeting bot join the meeting, similar to an AI notetaker. Detection results and controls are visible only to the host.
4. Does RealMeeting record meetings or store personal data?
No. RealMeeting does not store recordings, voiceprints, transcripts, or personal data. Detection happens in real time only.
5. Do I need to do anything once RealMeeting is enabled?
No. Once enabled, RealMeeting runs automatically. After a scan completes, you’ll see whether the participant is labeled Authentic or Manipulated.
6. What should I say if participants ask about the RealMeeting bot?
You can keep your response brief and neutral. For example:
“It’s a security tool we use for sensitive meetings.”
“It helps verify participant authenticity during high-trust calls.”
You don’t need to explain how the tool works or discuss any results. If needed, you can say your security team manages it and follow up after the meeting.