The same guide, with a face. What Alias avatars do, where they come from, and what they deliberately do not do.
Maintained by Alias. Last updated 2026-09-08.
An Alias guide can carry an avatar, though it is optional and the same guide runs with or without one. Alias offers premium 2D and 3D avatars that give a guide a face and a voice, so a fast answer also feels like a warm one. The avatar is optional and the same guide runs with or without one, but where it is used it changes the character of the interaction from reading a reply to being spoken to by an on-brand host. Avatars come from two routes. An organisation can pick one from a catalogue of premium avatars, available through Alias's partnership with HTC Viverse. Or it can have one made, modelled on its brand, its own staff, or original characters it owns. Either way the avatar is white-labelled to the organisation rather than being a generic assistant face.
Source: Alias
Expression, gaze and posture follow the meaning of every answer, not just the words. That distinction matters. A system driven only by phonemes produces a mouth that moves in time with speech but a face that stays blank, which reads as uncanny. Alias drives the performance from the semantic content of the response, so an avatar delivering bad news, a caveat, or a straightforward confirmation carries a different expression and posture in each case. The animation is AI-driven from the text of the response. There is no live motion capture behind it and no performer in the loop, which is what makes it viable at the scale of every visitor conversation rather than only for produced video.
Source: Alias
An organisation can choose an Alias guide's voice from a selection or have one made. An organisation can choose a voice from Alias's selection or have one made, and Alias's stated aim is the same warmth in every language the guide serves. This is the part that is easy to underrate. A guide answering across more than 100 languages with a single flat synthetic voice sounds like a machine translation; a consistent voice character across languages is what makes the guide read as one host rather than as many. Voice output is generated from the text of the response by third-party speech synthesis providers, which Alias lists publicly in its sub-processor list at alias.cm/legal/sub-processors.
Source: Alias
Alias avatars do not use a camera, a microphone or facial tracking, and Alias states this explicitly. It does not use cameras, microphones or facial tracking to animate avatars or to identify users. Where a guide uses a synthetic voice or an animated avatar, that output is generated from the text of the response, not from any live input from the visitor. This is a deliberate design choice rather than a limitation. It means a visitor gets an animated, expressive host without granting camera or microphone access, without an account, and without being identified. Where a guide produces synthetic audio or animated video, Alias marks that output as machine-readable artificially generated where technically feasible, in line with its Article 50 obligations under the EU AI Act.
Source: Alias privacy policy