Case Studies / XPOMET Medicinale 2024
HEALTHTECH · EVENTS & CONVERSATIONAL AI
Multilingual AI avatar replacing manual visitor guidance at a live healthcare conference
At XPOMET Medicinale 2024 in Leipzig, we deployed Paula – a voice-activated AI avatar that handled visitor questions, navigation, and exhibitor recommendations in real time. Built in partnership with Johannesstift Diakonie, Paula replaced the need for dedicated information staff in a noisy, high-traffic conference environment.
Client: Johannesstift Diakonie Services GmbH at XPOMET Medicinale 2024 – one of Europe’s leading healthcare innovation conferences, held in Leipzig.
KEY RESULTS
2 langs
Automatic English/German detection and mid-conversation switching
Live
Real-time voice interaction in noisy conference environment
Local
Wake-word detection processed on-device – no cloud dependency
Web+DB
Live web search and event database access for real-time answers
INDUSTRY
Health-Tech / Events
USE CASE
AI avatar for visitor guidance
AI APPROACH
LLM + speech + function calling
LANGUAGES
English & German (auto-detect)
ENVIRONMENT
Live conference, high noise
PARTNER
Johannesstift Diakonie
The challenge
XPOMET Medicinale is a large-scale healthcare innovation conference with dozens of exhibitors, a packed presentation schedule, and a diverse international audience. Attendees needed to find specific booths, check schedules, and get information about companies and sessions – all while navigating a noisy, crowded event space.
Traditionally, this requires dedicated information staff stationed at help desks – people who need to know the full event schedule, exhibitor details, and venue layout. Staffing this is expensive, inconsistent, and doesn’t scale to handle peak traffic when everyone arrives at once or sessions change.
The brief: build an AI-powered interactive assistant that could handle visitor questions in real time, in multiple languages, in a live conference setting – replacing the need for manual information desk staffing.
What we built
We built and deployed Paula – a voice-activated AI avatar stationed in the Pauls Health Area at the conference. Paula wasn’t a simple chatbot on a screen. She was a full conversational AI system handling natural spoken dialogue with conference visitors.
Automatic language detection and switching. Paula detected whether a visitor was speaking English or German and responded accordingly – no manual language selection needed. Visitors could even switch languages mid-conversation, and Paula followed seamlessly.
Wake-word activation with local processing. Interactions started with a simple “Hallo Paula.” The wake-word detection model ran locally on the device – no audio data sent to the cloud for activation. Paula also detected when a conversation had ended and deactivated automatically.
Event knowledge and function calling. Paula had access to the full event database – exhibitor details, the conference program, session times, and booth locations. Using LLM function calling, she could answer specific questions like “Which companies are presenting on digital pathology?” or “When is the next panel on AI in diagnostics?” and provide personalized recommendations based on visitor interests.
Dynamic visual display. During conversations, Paula identified visitor intent and displayed relevant exhibitor logos and additional information on screen – turning the interaction into a visual navigation experience, not just a voice exchange.
Real-time web search. Beyond event-specific questions, Paula could answer general knowledge queries using live web search – anything from local weather to factual questions – making her useful beyond just conference logistics.
The results
BEFORE
Visitor guidance handled by dedicated information desk staff. Limited capacity during peak hours. Single-language interactions. No personalized recommendations.
AFTER
AI avatar handling unlimited concurrent visitor questions in two languages. Personalized exhibitor recommendations. Visual navigation assistance. No staffing dependency.
Paula handled visitor interactions throughout the event, providing consistent, accurate information regardless of traffic volume. Attendees could get answers immediately without waiting in line or searching through printed programs. The multilingual capability was particularly well received by the international audience.
For exhibitors, Paula served as an additional discovery channel – actively directing visitors to relevant booths and sessions based on their stated interests, increasing visibility for participating companies.
The deployment demonstrated that conversational AI avatars can work reliably in challenging real-world conditions – not just quiet demo environments, but noisy, high-traffic live events with unpredictable user behavior.
Technology used
More Case Studies
See how we’ve helped other companies
AUTOMOTIVE (UNDER NDA)
Replacing reactive maintenance with a predictive maintenance roadmap
Unplanned equipment failures were causing costly production stops. We assessed the manufacturer’s infrastructure and delivered a concrete architecture for predictive maintenance, a roadmap from reactive repairs to data-driven prevention.
PdM
Architecture
Roadmap
Delivered
MANUFACTURING · RADAWAY
Making email-based order processing reliable with LLMs
Staff were manually reading customer emails, identifying products, and entering orders by hand. We turned a promising AI prototype into a production system that handles it end to end, across languages, formats, and attachments.
-90%
Manual intervention
95%+
Match accuracy
LOGISTICS · FR. MEYER’S SOHN
Eliminating manual data extraction from thousands of daily shipping emails
Operations staff were manually reading German and English logistics emails to pull out routing and scheduling data, every single day. We built an AI pipeline that extracts, structures, and delivers the data automatically.
–80%
Manual effort
2 langs
DE & EN
On-prem
Deployed