CTI needed a scalable way to assess pronunciation and support Chinese language learners worldwide without relying only on human evaluators and inconsistent feedback loops.
Reported impact
CTI expanded AI-supported Chinese language assessment globally, improving evaluation consistency and learner feedback while saving material, labor, and instructional effort.
Chinese Testing International Company Limited built its Learn Pronunciation to HSK application on Azure AI Speech to give Chinese language learners immediate pronunciation assessment at global scale. Learners speak into the app, Azure AI Speech analyzes pronunciation quality, and the system returns targeted feedback plus follow-on practice content that closes the loop across teaching, learning, testing, and assessment. CTI designed the application to support multiple difficulty levels and repeated practice cycles, while its broader roadmap also ties AI into facial recognition, test recognition, and intelligent assessment workflows. The architecture turns what used to require scarce expert evaluation into a consistent cloud service that can reach a worldwide learner base. For CTI, the result is a mobile pronunciation classroom that supports 1,330 test centers across 162 countries with more reliable evaluation, faster feedback, and lower demand on human grading resources.
Learner speech is captured in Learn Pronunciation to HSK, analyzed by Azure AI Speech for pronunciation quality, and then returned as targeted assessment feedback and repeated practice content that reinforces weak areas.
Presently, CTI has set up 1,330 test centers across 162 countries, catering to over 50 million Chinese learners worldwide.
The application leverages Azure AI Speech pronunciation assessment to provide instant, accurate, and targeted pronunciation assessment feedback to Chinese language learners around the world.
Using Azure AI Speech Services, CTI created the Learn Pronunciation to HSK application, providing instant pronunciation feedback to learners.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Microsoft Sample Oct 15, 2026 Bar-Ilan University Students who miss lectures struggle to reconnect new lessons with prior material when course information is spread across notes, recordings, slides, and readings, and lecturers need a reliable way to answer routine… 71 courses Azure AI Services · Microsoft Azure · Microsoft Copilot Studio Microsoft Sample Jan 27, 2025 The University of Hong Kong HKU faced difficulties with a complex onboarding process for its diverse student body and increasing administrative workload for faculty, necessitating a more efficient system for information delivery and workflow… The University of Hong Kong leverages Microsoft 365 Copilot and Copilot Studio to build a more dynamic learning… Microsoft 365 · Microsoft 365 Copilot · Microsoft Copilot Microsoft Sample Dec 13, 2024 Universitas Terbuka Indonesia's Universitas Terbuka (UT), with over 525,000 active students and growing, struggled to provide timely tutor support, risking student engagement, satisfaction, and retention. The AI tutor enhances student grades and engagement by providing more comprehensive support 24 hours a day. Azure AI Document Intelligence · Azure AI Foundry · Azure Cosmos DB 