Definition: An avatar is the digital representation of a person in a virtual or augmented environment. It ranges from a simple head-and-hands model to a photorealistic representation that transmits gaze, facial expressions and gestures in real time. In VR and mixed reality applications, the avatar shows who is present, where someone looks and what someone points at.
In B2B mechanical engineering, companies use avatars for: VR meetings with distributed teams, joint design reviews on the 3D model, remote support between an expert and a technician, training with a trainer in the virtual space, and sales appointments in the virtual showroom.
Advantage: The participants share one space. Pointing, gaze direction and distance replace complicated descriptions of the screen.
English: Avatar
German: Avatar, digitale Stellvertreterfigur
How It Works
From the sensor to the figure
A real-time avatar is created in four steps:
- Capture: The headset measures the head position and rotation. Controllers or hand tracking supply the hand movements
- Extended signals: Eye tracking and face tracking capture gaze and facial expressions (depending on the device)
- Reconstruction: Software adds the body parts that are not measured (for example elbows, legs) with inverse kinematics
- Transmission: Only motion data goes to the other participants, not the video
Result: Low bandwidth, but body language you can still recognize.
Avatar types
| Type | Representation | Typical use |
|---|---|---|
| Stylized | Cartoon-like, head and hands | Quick meetings, trade show demos |
| Semi-realistic | Full body, individual appearance | Regular team meetings |
| Photorealistic | Generated from a face scan | Customer appointments, conversations where presence is critical |
| Volumetric | Recording of a real person | Pre-produced expert videos |
Photorealistic avatars
On the Vision Pro, Apple generates a "Persona" from a face scan. The sensors of the headset control it in real time. Meta has researched "Codec Avatars" for years. Until now, they are a research project.
B2B Applications
Design review: together at the machine
Scenario: Design engineers in Germany, a customer in the USA and a supplier in Italy examine a new packaging machine at a 1:1 scale.
Avatar benefits: - Everyone sees who stands where and points at what - Discussion at the actual assembly instead of on screenshots - You can experience ergonomics questions (reach, lines of sight)
Result: Fewer misunderstandings, less travel.
Remote support in service
A technician stands on site at a CNC milling machine. The expert sits in the factory:
- The expert appears as an avatar in the AR view of the technician, or both meet at the digital twin
- Pointing gestures mark the correct component
- The expert demonstrates the work steps in the model
Training with a trainer avatar
Participants at several locations train together on a virtual filling line. The trainer sees who works where and steps in. Alternative: A pre-recorded trainer avatar guides the participants through standard procedures.
Sales and trade shows
In the virtual showroom, the sales representative accompanies the customer as an avatar. At the trade show, a colleague from headquarters can join the conversation through VR.
Procedure of a sales appointment: 1. The customer gets a link (browser) or a prepared headset 2. Sales and the customer meet as avatars in front of the machine 3. Sales opens assemblies and shows options and dimensions 4. Questions are answered directly on the model, and notes stay saved
Advantage: A specialist from engineering can join at short notice without travel. On a kiosk at the booth, you can show the same scene to visitors without a headset.
Remote collaboration in daily work
For short alignments without a 3D reference, a video conference usually stays more efficient. Avatars are worth it where the spatial relation to the object is important.
Avatars in Practice
Platform landscape
The market for VR meetings is changing. Microsoft discontinued Mesh at the end of 2025 and integrated avatar-based 3D events into Teams. Meta shut down Horizon Workrooms in February 2026. Consequence for companies: Pay attention to open formats and interchangeable tools.
Cross-platform participation
Not every participant has a headset. Good solutions allow mixed groups: VR users as full-body avatars, desktop or tablet users with a simple figure or a video tile.
Challenges
- Uncanny valley: Almost realistic figures quickly look strange. Users often accept stylized avatars better
- Partial capture: The software usually estimates the legs and lower body. It does not measure them
- Biometric data: Face scans and facial expression data are personal data (data privacy)
- Identity: In a customer appointment, it must be clear who is behind the avatar (data security)
Presence and Communication
What avatars can do that video cannot
In a video conference, all participants see the same image, but nobody shares a space. An avatar, on the other hand, has a position in the virtual space. This gives signals that make conversations at the 3D model easier:
- Spatial sound: Voices come from the direction of the speaker
- Gaze direction: You can see who looks at which component
- Pointing: A hand gesture replaces the description "top left, the second valve"
- Distance: Small groups form naturally, for example at different stations of a line
Introduction in the team
For avatar meetings to work in daily work, companies should clarify some points in advance:
- Define the use case: for example a weekly design review instead of general meetings
- Supply the hardware: Prepare the headsets, charge them, keep the firmware up to date
- Instruction: A short introduction to navigation and gestures before customers take part
- Allow hybrid participation: Include colleagues without a headset through desktop or tablet
- Prepare the data: Load and check the 3D model before the appointment
Session duration
Long VR sessions are tiring. For reviews and training, shorter blocks with breaks work well.
Frequently Asked Questions (FAQ)
Does a B2B meeting need photorealistic avatars?
Not necessarily. For design reviews, the head, hands and pointing gestures are sufficient, because the object is the focus. For customer conversations, realism can increase trust. Rule of thumb: The more important facial expressions are, the more realism is worth it.
Which hardware is necessary?
Minimum: A standalone VR headset with hand tracking. For facial expressions: Headsets with eye and face tracking. Without a headset: Desktop or tablet access with a simplified representation.
How are avatars created from face data?
The headset or a smartphone scans the face, and software generates a 3D model from it. During use, the sensors control gaze and facial expressions. Data privacy: Clarify the storage location and purpose of this biometric data before the rollout.
Is there an open avatar format?
VRM is an open format for humanoid avatars based on glTF 2.0. Since 2024, the VRM Consortium and the Khronos Group have worked on its standardization as glTF extensions. In the enterprise environment, proprietary formats are still common.
What is the difference between an avatar and volumetric video?
Sensor data controls an avatar live. Volumetric video is a recording of a real person that plays back later. It is realistic, but not interactive.
Industry Standards and Technologies
- Khronos OpenXR – open interface for headsets, hand tracking and eye tracking
- VRM Consortium – open avatar format based on glTF
- Khronos glTF – basis for avatars and 3D models
- Metaverse Standards Forum – working groups on interoperable avatars
- W3C WebXR Device API – avatars and VR in the browser
- Apple visionOS Developer – Personas and SharePlay in visionOS
Related Topics
XR basics: - Virtual reality – fully virtual environment - Augmented reality – digital content in a real environment - Mixed reality – merging of both worlds - Extended reality (XR) – umbrella term - Immersion – feeling of being immersed
Tracking: - Eye tracking – gaze tracking - Sensors – motion capture - Markerless tracking – tracking without markers - Hand tracking – capture of the hands - Face tracking – capture of facial expressions
Devices and platforms: - visionOS – Apple spatial computing - HoloLens – Microsoft MR headset - WebGL – 3D in the browser - OpenXR – open XR interface - VRM – open avatar format
B2B applications: - Digital twin – digital replica - Immersive experience – immersive presentation - Remote collaboration – distributed collaboration - Design review – examination on the 3D model - Metaverse – persistent 3D spaces - Volumetric video – recording of real people
