Embedded Voice Assistant for VETC App
Tasco · Posted 1 year ago
Automotive and mobilityConversational AI / ChatbotsVoice AISearch & Knowledge Retrieval
- Budget
- To be determined
- Timeline
- Not specified
- Closed on
- 31 Oct 2025, 23:59 ICT
Problem statement
VETC’s, Vietnam's leading electronic toll collection (ETC) system operator mobile app currently requires users to navigate multiple menus and screens for basic tasks, which is cumbersome—especially while driving or in urgent situations—leading to frustration and reduced engagement.
Expected outcomes and success metrics
Enable hands-free, natural-language access to core VETC features (wallet management, service locator, etc.), with high ASR accuracy, low latency, and contextual understanding in Vietnamese.
Current process
Evaluating VNG Kiki, VinBigData Vivi, FPT SmartCloud
Current solution
Evaluating VNG Kiki, VinBigData Vivi, FPT SmartCloud
Target users and impacted teams
Over 2 million VETC app users across Vietnam
Data availability and readiness
Real-time GPS location, wallet balance API, service station directory; mock APIs available for PoC
Integration, deployment and infrastructure
Wake-word activation (“VETC ơi”), Vietnamese ASR with <10% WER, contextual dialogue engine, on-device or lightweight voice model (<200MB), integration with VETC APIs
Relevant AI technologies
Conversational AI / ChatbotsVoice AISearch & Knowledge Retrieval