Bidirectional real-time translation for live Hebrew and English conversations
Generative AI · Service desk
Real-time Language Translation for an Enterprise Service Desk
A secure translation and agent-assist platform removed language friction from live support while improving response quality with contextual knowledge recommendations.
About the customer
Technology services
A global technology solutions company with more than 150 years of industry heritage and a UK service desk of over 120 agents supporting Hebrew-speaking customers in Israel.
Challenge
Removing the barriers to a connected experience
- Hebrew-speaking customers and English-speaking service agents lacked a real-time bilingual support experience.
- The service desk needed accurate communication across many concurrent live-chat sessions.
- Structural and semantic differences between Hebrew and English increased miscommunication risk.
- Agents lacked contextual knowledge suggestions and AI-assisted responses inside their existing workflow.
Approach
An enterprise-ready AI solution
Quantaleap embedded bidirectional Hebrew–English translation into the existing chat experience with minimal workflow disruption.
A domain-adapted language model handles Hebrew linguistic nuance, while a RAG layer retrieves relevant knowledge and recommends contextual responses.
The enterprise architecture combines a responsive presentation layer, low-latency APIs, vector and relational data services, role-based identity and containerized elastic deployment.
Project highlights
What the solution delivered
Fine-tuned DictaLM 2.0 model designed for Hebrew linguistic nuance
RAG-powered agent assistance using semantic knowledge retrieval
Elastic scaling for parallel customer conversations
Enterprise authentication and role-based access for more than 120 agents
Measured outcomes