If you are a city or national migration agency struggling with long processing queues and language barriers when receiving newcomers — this project developed a multi-agent AI platform with personalized multilingual conversation agents that guide migrants through registration and orientation. It was validated across 3 pilot use cases in different European contexts with a consortium of 16 partners, reducing the burden on caseworkers while improving migrant outcomes.
AI-Powered Multilingual Assistants That Help Governments Integrate Migrants Faster
Imagine a newcomer arriving in a country where they don't speak the language, don't know the rules, and have no one to guide them through the bureaucracy. WELCOME built smart virtual assistants — think of a patient, multilingual guide on your phone — that walk migrants through registration, language learning, and civic education step by step. These AI agents adapt to each person's background and language, and for the government side, there's a dashboard that helps case workers see the big picture and make better decisions. The whole thing was tested in real pilot programs across three European locations.
What needed solving
Governments and municipalities across Europe face an overwhelming challenge: millions of migrants need to be registered, oriented, taught the local language, and integrated into society — all while dealing with language barriers, cultural differences, and overworked caseworkers. Current systems rely on one-size-fits-all information portals that don't adapt to individual needs. The result is slower integration, higher costs, and worse outcomes for both migrants and host communities.
What was built
WELCOME built a multi-agent AI platform with three progressive prototypes and a final demonstrator. The platform includes personalized multilingual conversation agents for migrant-facing services (registration, orientation, language teaching, civic education), immersive VR/AR technologies for realistic integration experiences, and decision support dashboards with visual analytics for public administrators. All validated across 3 European pilot sites.
Who needs this
Who can put this to work
If you are an EdTech company looking to serve the growing migrant language-learning market — WELCOME built immersive VR/AR-based language teaching modules with embodied conversation agents that adapt to each learner's profile and native language. The system handles multiple languages and was developed with EUR 3,995,709 in EU funding, producing 21 deliverables including a final working demonstrator.
If you are a migrant support organization overwhelmed by caseloads and struggling to provide personalized guidance — this project created decision support tools using visual analytics and semantic reasoning that help caseworkers track integration progress and allocate resources. The platform was built by 16 partners across 5 countries, including 6 entities directly involved in migrant reception and integration.
Quick answers
What would it cost to license or deploy this platform?
The project received EUR 3,995,709 in EU funding as an Innovation Action, which typically means the technology is closer to market than basic research. Specific licensing terms are not publicly available — you would need to contact the coordinator at Universidad Pompeu Fabra. Based on available project data, the platform is modular, so costs would depend on which components you deploy.
Can this scale to handle large migrant populations across multiple countries?
The system was specifically designed for portability and validated in 3 pilot use cases across different European contexts. The consortium spans 5 countries (Czech Republic, Germany, Greece, Spain, UK), which suggests the platform can adapt to different regulatory and linguistic environments. The multilingual capabilities are built into the core architecture.
Who owns the intellectual property and can I license individual components?
IP is typically shared among the 16 consortium partners according to their EU grant agreement. Universidad Pompeu Fabra as coordinator would be the first point of contact. The modular multi-agent architecture suggests individual components (language teaching, decision support, conversation agents) could potentially be licensed separately.
Does this comply with EU data protection regulations for sensitive migrant data?
As an EU-funded project dealing with vulnerable populations, WELCOME would have been subject to strict ethical review and GDPR compliance requirements. The project involved 6 entities directly working in migrant reception, suggesting real-world data handling practices were built in. Specific compliance documentation should be requested from the coordinator.
How long would it take to deploy this in our municipality?
The project ran from February 2020 to April 2023 and produced 3 progressive prototypes culminating in a final demonstrator. Based on available project data, a deployment timeline would depend on language requirements and local system integration, but the 3 pilot validations suggest the platform can be adapted to new contexts.
Can this integrate with our existing case management systems?
The platform includes decision support technologies with visual analytics and semantic reasoning designed for public administration use. The multi-agent architecture was built to be service-oriented, which typically supports integration with existing IT infrastructure. Specific API and integration capabilities should be confirmed with the development partners.
Who built it
The WELCOME consortium of 16 partners across 5 countries brings a balanced mix for a technology entering the market: 3 ICT companies provide the technical backbone, 4 universities and 4 research institutions deliver the AI and language technology expertise, and critically, 6 entities are directly involved in migrant reception and integration — meaning the end users shaped the product from day one. With 4 SMEs in the mix and a 19% industry ratio, this is still heavily research-driven, but the inclusion of the International Organization for Migration as a subcontractor adds serious institutional credibility. For a business looking to partner or license, the coordinator Universidad Pompeu Fabra in Spain is the gateway, and the multi-country spread (CZ, DE, EL, ES, UK) means the platform has already been stress-tested across different regulatory environments.
- UNIVERSIDAD POMPEU FABRACoordinator · ES
- ETHNIKO KENTRO EREVNAS KAI TECHNOLOGIKIS ANAPTYXISparticipant · EL
- NTT DATA SPAIN, SLparticipant · ES
- NUROGAMES GMBHparticipant · DE
- PRAKSIS ASSOCIATIONparticipant · EL
- VYSOKE UCENI TECHNICKE V BRNEparticipant · CZ
- DEUTSCHES FORSCHUNGSZENTRUM FUR KUNSTLICHE INTELLIGENZ GMBHparticipant · DE
- SHEFFIELD HALLAM UNIVERSITYparticipant · UK
- KENTRO MELETON ASFALEIASparticipant · EL
- UNIVERZITA KARLOVAparticipant · CZ
Universidad Pompeu Fabra (Barcelona, Spain) — look for the WELCOME project lead in their Department of Information and Communication Technologies
Talk to the team behind this work.
Want to explore licensing WELCOME's AI-powered migrant integration tools for your municipality or organization? SciTransfer can arrange an introduction to the development team and help assess fit for your specific context.