Virtual host
A screen at your entrance or exhibition stand greets attendees, gives event information and answers questions on the spot. In use at DCK 2024 and XPOMET 2024.
AI Avatars · Conversations in real time
AI avatar development for conversations with your visitors and customers
theBlue.ai develops AI avatars for companies: virtual characters that talk with visitors, customers and employees in real time, by voice or text, and answer from your own data. They run on the web, on large screens or in virtual reality and speak the languages you need.
Proof
Avatars on busy event floors, in three numbers
Our avatars answered visitors live at two medical congresses in Leipzig in 2024, in noisy halls, by voice and in more than one language.
4 days
Continuous visitor support at the 141st German Congress of Surgery, Leipzig, April 2024
2 languages
English and German, detected automatically at XPOMET Medicinale 2024
Kiosk + VR
The same congress assistant on a large screen and on Meta Quest 3
Some of the organisations we have built for
Use cases
Where an AI avatar brings a conversation to life
From the entrance of your building to customer service and training. Each avatar is tailored to your industry and answers from your own knowledge.
Product specialist
On-demand product demonstrations and detailed answers for professionals who want to keep up with what is new.
Customer service
Immediate assistance around the clock across your service channels, resolving inquiries with precision.
Medical assistant
Supports patient data collection and gives professionals quick access to a large knowledge base.
Virtual HR assistant
Helps with onboarding new employees, delivers training programmes and gathers basic candidate information in pre-interviews.
Virtual trainer
Customised workouts and professional guidance at home or in any facility, to keep people motivated.
How it works
A conversation a visitor can simply start
- 1
The visitor speaks
No app, nothing to type. Wake-word detection on the device starts the conversation, as with “Hallo Paula” at XPOMET, and it ends by itself when the visitor walks away.
- 2
The avatar understands, even on a busy floor
Professional microphones and noise reduction keep speech-to-text reliable in loud halls. The avatar detects the language and follows a switch mid-conversation.
- 3
It answers from your data
A large language model with RAG answers from your documents or event database, function calling lets it look things up, and text-to-speech gives the answer a voice while the screen shows what the visitor needs.
From idea to launch
How your avatar comes to life
Five steps from the first idea to an avatar that talks with your visitors, with fast iterations along the way.
-
1
Concept
Use case and setting
Who talks to the avatar, what it should do, and where it runs: web, big screen or VR.
-
2
Languages
Your audience’s languages
The avatar speaks the languages your visitors need and answers in the one they use.
-
3
Your data
Knowledge connected
Documents, event data and tools are connected through RAG, in line with your privacy rules.
-
4
MVP
Fast iterations
A first working version on reliable frameworks, refined with real questions.
-
Go live
Talking with visitors
A fully operational avatar at your stand, desk, website or VR space.
Case studies
Avatars and assistants in live use
Voice assistants that guided visitors at two medical congresses, and a virtual assistant that answers from hundreds of product documents.
4 days
Continuous
100+
Sessions
Healthcare · German Congress of Surgery
A voice assistant answering attendee questions for four days
At Germany’s largest surgical congress, attendees asked staff the same navigation and scheduling questions all day. A voice assistant took those over.
Read the case study
2 languages
Detected automatically
Live
Voice interaction
HealthTech · Xpomet Medicinale 2024
An AI avatar handling repeat visitor questions at a congress
Visitors at a healthcare innovation congress kept asking the same questions about exhibitors, sessions and directions. An avatar answered them in English and German, so staff could hold the conversations that mattered.
Read the case study
RAG
Over their own files
Seconds
With source references
MedTech · apoQlar
A retrieval assistant that answers from hundreds of product documents
Support staff searched through hundreds of product documents to answer customer and compliance questions. They now ask in plain language and get an answer with the source it came from.
Read the case studyHow to start
Tell us where your avatar should talk
A trade fair stand, a service desk, a training programme or a VR space. We start with the use case and the data behind it.
Describe your use case
Where the avatar should run, who talks to it and what it needs to know. We come back within one business day with an initial assessment and a proposal for a short call.
Discuss your projectFAQ
Questions about AI avatars
An AI avatar is a virtual character powered by generative AI that holds a conversation with people in real time, by text or by voice, and answers from a company’s own data. It combines speech-to-text, a large language model with retrieval-augmented generation (RAG) and text-to-speech. theBlue.ai develops AI avatars as web assistants, as standalone applications on large screens and as 3D avatars for virtual reality.
A chatbot answers in text inside a chat window. An AI avatar adds a face and a voice: people talk to it out loud, it answers in speech, and it can show content on screen while it talks. That suits places where nobody wants to type, such as an event floor, a reception or a VR space. Both can answer from your own data.
On a website, on large screens and in virtual reality, for example on Microsoft Mesh. At the German Congress of Surgery the same assistant ran on a screen kiosk and as a VR experience on Meta Quest 3.
Yes. A congress floor is acoustically hostile, so we combine professional microphone hardware with software noise reduction, and process speech in parallel so the answers come back fast enough to feel like a conversation. That setup ran through four days of the German Congress of Surgery.
At the 141st Congress of the German Society of Surgery (DCK 2024) at the Leipzig Congress Centre from 23 to 26 April 2024, a voice assistant guided visitors through hundreds of parallel sessions on a screen kiosk and in VR. At XPOMET Medicinale 2024 in Leipzig, the avatar Paula, built with Johannesstift Diakonie, answered visitor questions in English and German. See the DCK case study and the XPOMET case study.
The languages you want. The avatar speaks the languages your visitors and customers need, and it can detect which one a person is using and answer in it. At XPOMET Medicinale our avatar Paula recognised English and German automatically and followed visitors who switched language mid-conversation.
It answers from your company data through retrieval-augmented generation, such as documents or an event database with exhibitors, sessions and locations. Through tools it can look things up and perform actions, and it can recommend sessions or show relevant content on screen while it talks.
We build every avatar to comply with your privacy regulations, so your information stays secure. At XPOMET the wake-word model ran locally on the device, which meant no audio went to the cloud just to start a conversation.
We guide you from the initial concept through an MVP to a fully operational avatar, with fast iterations built on reliable technologies and frameworks. Each project is a one-time service that ends with a fully functional avatar.
Ready to bring AI to life for your customers?
Tell us where your avatar should talk and what it needs to know. We’ll come back within one business day with an initial assessment.
- Senior engineers from day one. No account managers, no handoffs.
- Your avatar answers from your own data and follows your privacy rules.