We built an AI support agent trained on Greenwich's enrolment policies, student handbook, and course catalogue. It handles transfers, course progress, fees, document requests, and welfare cases — with humans in the loop on regulated decisions.
This agent is loaded with three demo students. Click any prompt below to copy it — then paste into the chat widget or call the number to test on voice.
Three demo students: Priya Sharma (GH238472), Carlos Rodriguez (GH912038), Aarav Patel (GH551207). Each has its own attendance, course progress, fee status, age, and visa profile. Make up any other details — the agent rolls with whatever you give it.
The same workflows you can test in the chat widget also run on voice — with one-question-per-turn pacing, spell-back on names and emails, and headline-only responses. Call the demo number from any phone.
0468 024 850+61 4 6802 4850 from outside Australia
AI support that understands CRICOS regulation, visa-impacting decisions, multi-campus operations, and the full range of student services enquiries.
Standard 7 best-interests assessment for under-6-month transfers. ELICOS attendance vs VET progress thresholds. 28-day visa rule on cancellations. The agent applies the right rule for each student.
The agent never auto-approves a transfer, deferral, suspension, or cancellation. It gathers the right information, submits a structured request, and authorised Student Services staff make the call.
Transcripts, CoE letters, Letters of Offer, release letters, statements of attainment, status letters for visa applications. Each tailored to the recipient.
Distress, fear of expulsion, self-harm ideation — the agent surfaces Lifeline 13 11 14 and UPRISE 1300 209 371, then warm-hands off to a Student Support Officer at the right campus.
Sydney, Melbourne, Brisbane, Gold Coast, Perth, Adelaide, plus the Greenwich Online Academy. The agent routes handoffs to the right campus Student Services inbox.
Workflows are built once and deployed across chat, voice, email, and SMS. Out-of-hours phone enquiries get the same accurate answer as a business-hours chat.
Grounded in real Greenwich source material. No fabricated policies, no invented payment instructions, no made-up intake dates.
Course catalogue, student handbook, six enrolment policies, and six official forms — split into workflow-shaped instructions and knowledge-shaped content.
Twelve specialised workflows covering transfer, course progress, defer/suspend/cancel, fees, credit transfer, transcripts, visa conditions, welfare, class change, document requests, login support, and FAQ.
Student lookup, attendance, course progress, timetable, plus seven write tools that create HITL tasks. In production these connect to your Student Management System, Canvas, and PRISMS.
No auto-confirming visa-impacting decisions, no migration advice, under-18 parent-consent gate, welfare-crisis routing. Every outgoing message is checked.
From this demo to production in weeks, not months. Phased so the team builds confidence as it scales.
Connect to HubSpot Conversations and your existing MCP endpoints for the Student Management System and Canvas. Replace mock student records with real lookups. Keep the policy knowledge base.
The agent answers, looks up, and drafts — humans approve any regulated action. You see how the agent performs against the real ticket queue before write-back goes live.
The agent submits HITL requests directly. Humans stay in the loop on regulated decisions — Standard 7 assessments, intervention strategies, deferral approvals. Data entry is gone, decision-making isn't.
Same agent, same knowledge, on chat, voice, email, and SMS. After-hours phone enquiries get covered. Outbound flows reach the right student at the right moment.
Routine enquiries — timetable, document requests, fee questions, work-hour rules — resolve without a human touching them.
Your Student Services Managers and Academic Managers spend their time on the regulated decisions that need their expertise — not on data entry.
Coach lets a non-technical CS lead interrogate why the agent said something, correct knowledge in plain English, and add new workflows over time.
Configurable guardrails check every outgoing message for hallucination, absolutes, and migration advice. Every regulated decision keeps a paper trail.
Ticket Quality Score (independent of CSAT) lets you benchmark agent quality against your human team on factual accuracy and instruction-following.
Short sentences, no jargon, no sycophantic openers, no clinical advice. Lifeline and UPRISE surfaced when needed.
We'd love to walk you through how this works with your real data and integrations.
Talk to the team