news
SAMI: AI-Powered Chatbot for Migrants' Information and Humanitarian Data Collection
Developed by the Mixed Migration Centre (MMC) and DRC, SAMI is an AI-powered WhatsApp chatbot that delivers tailored, verified information to migrants on rights, services, and humanitarian assistance. By gathering anonymised feedback and insights, it helps identify emerging needs among people on the move. Following a successful pilot in Colombia in 2025, SAMI is now expanding to scale.
The challenge
Across Latin America, West Africa, and beyond, people on the move navigate some of the world's most dangerous migration corridors guided only by informal WhatsApp groups, advice from smugglers, and rumours. Migrants making critical decision — about routes, legal options, shelter, and safety — based on unverified information exposes them to exploitation, legal uncertainty, trafficking, and life-threatening risks. At the same time, humanitarian actors and public institutions lack cost-effective, real-time tools to understand shifting protection needs, counter harmful narratives, or adapt their programming in rapidly changing contexts.
The opportunity
SAMI was designed to address these challenges simultaneously: a single tool that empowers migrants with reliable, contextualised information while generating actionable evidence for the humanitarian system. SAMI is a generative AI chatbot operating exclusively through WhatsApp — the most widely used messaging platform among migrants globally. It draws on a curated, expert-validated knowledge base to answer users' questions in real time. Critically, SAMI does not have access to the open internet: all responses are generated solely from the verified MMC/DRC database, removing the hallucination risks inherent to general-purpose AI tools.
Fostering innovation
SAMI is an innovative tool to respond to an urgent challenge. SAMI is powered exclusively by a consolidated and verified database developed by MMC and DRC Colombia, strengthening the link between data-driven research and operational realities in the field. It does not perform internet searches, ensuring that its responses are based solely on previously validated and updated information. It also facilitates continuous learning to inform humanitarian response, as it operates a consent-based, anonymous and ethical real time data collection function.
More content like this
news
Mainstreaming Environmental Approach in DRC's Livelihoods Strategy
news
Mixed movements research through social media listening in the LAC region
news
Digital inclusion and access to decent livelihood opportunities
news
Using The DEEP Platform for Qualitative Protection Analysis to Design Better Evidence-based Responses in LAC
news
Press release: DRC commends Colombia’s decision to regularize Venezuelan migrants
news
Protection Monitoring Dash Board
news