Staff repeatedly answered common questions about courses, times and lesson arrangements
EDUCATION
AI support, class allocation, reminders and collection in one admin workflow.
A Hong Kong education provider wanted to reduce repetitive enquiries and admin follow-up so staff could focus on student cases requiring human judgment.
Client context
The client is a Hong Kong education provider handling course enquiries, class allocation, lesson reminders and outstanding payment follow-up. The work was spread across different staff and tools, which made unresolved cases difficult to track.
Original workflow
Class allocation, lesson reminders and outstanding payment follow-up were split across manual steps
Management lacked a clear view of unresolved cases
BOTTLENECK
The issue was not one difficult task. It was the cumulative cost of repeat questions and fragmented admin, forcing staff to switch context constantly.
What was built?
An AI chatbot handled common questions using approved answers
Student data entered rule-based allocation and reminder flows
Outstanding payments and exceptions were routed to named staff
Each status change was logged for team visibility
Results and measurement basis
Cases are anonymised and every metric includes its measurement basis rather than presenting a claim without context.
estimated admin hours saved weekly
Based on the client's previous workload estimate and cross-checked against system handling volume.
repeat enquiries handled by AI
Calculated from repeat enquiries completed without human handoff.
typical response time
Time to begin responding to an in-scope knowledge-base enquiry.
What changed after launch
AI answers common course enquiries immediately, reducing repetitive responses from staff
Class allocation, reminders and payment follow-up now run through one workflow, reducing missed follow-ups
Management can see the status of each case, making team handovers clearer
EXPERT AUTOMATION CONSULTATION
Put your most complex workflow problems in experienced hands.
Clarify the workflow, system constraints and next decision before choosing what to solve first.