Project Background
Scaling Intelligent Enquiry Handling Across a High-Volume Education Platform
HKU SPACE operates one of Hong Kong’s largest continuing education ecosystems, serving diverse learner segments including working professionals, secondary school graduates, and lifelong learners.
With thousands of programmes and frequent public enquiries, the institution required an AI-powered conversational layer capable of delivering accurate, contextual, and scalable responses — while remaining fully aligned with institutional governance and data control.
The objective was not simply to deploy a chatbot, but to architect an enterprise-grade AI agent framework deeply integrated with HKU SPACE’s digital platform.
Our Approach
AI-Powered Conversational Architecture
Designed and deployed a GenAI-based chatbot capable of handling open-ended enquiries, contextual follow-ups, and meaning-based programme search — delivering natural, dynamic and human-like interactions beyond rule-based automation.
Domain-Specific AI Agent Framework
Developed specialised AI agents tailored to HKU SPACE use cases, including Programme Advisor Agent and Event Advisor Agent, enabling precise, context-aware responses aligned with academic structures and programme attributes.
Seamless Integration with CMS
Fully integrated the chatbot within CMSCore, eliminating duplicate data entry and ensuring real-time synchronisation of structured content. All updates within the CMS are automatically reflected in AI responses.
See How It Works
Specialised AI Agents Behind the Chatbot Experience
Within the chatbot, domain-specific agents handle different types of enquiries — from institutional information to programme discovery and event guidance. Grounded in structured content and data, they work together to deliver consistent answers, context-aware recommendations, and guidance aligned with user intent.
Handles institutional and administrative questions, including admissions, campus information, and website content. Responses are grounded in structured CMS content to ensure consistency with official published information.
Reliable Responses Through Structured Guardrails
The conversational engine operates within structured guardrails to ensure accuracy and institutional control. Responses are generated from verified content, with contextual validation and controlled prioritisation applied before delivery. This layered safeguard framework reduces the risk of unsupported outputs while preserving natural language flexibility.
Enterprise-Grade Security & Control
The chatbot is designed with enterprise-grade security controls, including end-to-end encrypted communication, rate limiting, and managed API access to ensure system reliability. Bot and spam detection mechanisms protect against malicious activity, while all data remains within the HKU SPACE ecosystem, preserving institutional ownership and eliminating reliance on unmanaged external repositories.
