USCIS's bilingual virtual assistant 'Emma' answers about 456,000 immigration questions a month — correctly handling ~90% in English and ~86% in Spanish — easing pressure from over a million monthly calls.
~456,000
Questions answered per month
~90%
Correct resolution rate (English)
~86%
Correct resolution rate (Spanish)
>1 million
Monthly calls to USCIS (context)
Details
Maturity
Established
Promoter
U.S. Citizenship and Immigration Services (USCIS)
Keywords
Service delivery, bilingual, chatbot, immigration services, virtual assistant
Context
Emma is the virtual assistant on the USCIS website, named after poet Emma Lazarus, that helps people find immigration information and navigate the site in English and Spanish; it was introduced to relieve the load of over a million monthly calls to USCIS by offering 24/7 self-service.
Objectives
Help users find immigration information and navigate USCIS's website bilingually, reducing call-centre demand.
Activities
An automated Q&A and navigation assistant embedded on the USCIS website, answering common questions in English and Spanish.
Results
Emma answers an average of around 456,000 questions per month, correctly resolving roughly 90% of common questions in English and 86% in Spanish, and is used nearly twice as much as the site's search function.
Conclusions
Described as a solid, dependable example of bilingual self-service at a high-demand agency — effective and well-used, if more modest in scope and novelty than generative-AI assistants.
Implementation
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Data sources
Where this practice's information was retrieved from, and when.
Estonia's open-source virtual-assistant ecosystem giving citizens a single conversational channel to public services, integrated across 18+ organisations, with human hand-off …