Up Education logo.
Nursing students.

Every student question answered 24/7, on channels they trust.

Problem

Students generate a constant stream of everyday enquiries: where's the library? What's my timetable? How do I enrol in this module? These consume a disproportionate amount of time from staff, who are already under pressure to keep on top of increasing workloads. International cohorts with language and time zone differences add further friction.

At the same time, colleges have a duty of care to identify students who may be struggling, whether that's a student asking the same question over and over (potentially due to confusion or stress), or a genuine crisis situation buried in casual conversation. Relying on staff to catch these signals manually, across thousands of interactions, is neither scalable nor reliable.

Solution

We built an AI-powered student enquiry tool using RAG on GPT-5.5, orchestrated with LangGraph, that can answer almost any student question and translates seamlessly into any language. It's accessible wherever students already are, on the channels they’re familiar with, whether that’s the college website, WhatsApp or WeChat.

Beyond answering questions, the tool provides pastoral care through built-in sentiment mapping. It's designed to recognise urgency and escalate appropriately; a genuine safety concern is flagged to the college as a P1 immediately, while a student asking repetitive questions might instead be gently offered a human touchpoint and connected with their tutor Critically, the college, not the chatbot, decides what the tool is allowed to help with, and use cases are only extended deliberately, based on real usage data.

MakerTech purple 3D crop 2.

Result

The tool has proven highly successful, now supporting around 7,000 conversations a month, freeing up significant staff time that would otherwise go into answering routine enquiries. As it learns where it's genuinely useful, colleges have a clear, data-backed path to extend its scope over time. For example, into practical, non-academic questions like local nightlife or transport, without losing control over what the assistant is trusted to handle.

Just as importantly, the pastoral care layer gives colleges an early warning system they didn't have before, surfacing at-risk students and disengagement patterns that would otherwise go unnoticed and could lead to poor academic performance, student drop-outs or in fringe cases, disruptive behaviour on campus.

MakerTech purple 3D crop 3.

Technologies

LangGraph logo.
OpenAI logo.
Python logo.
WhatsApp Business API logo.