The Student Support AI

The numbers tell a clear story. Student-to-counsellor ratios at NZ tertiary institutions have stretched well beyond recommended levels. Wait times for academic advising can extend to weeks during peak periods. And the students most at risk of dropping out are disproportionately those who face the most barriers to accessing in-person support: part-time students, distance learners, students with work commitments, and students who simply do not know where to start.

Josiah has spent two decades in education technology. His work at Western Sydney University took online course completion from 40% to 100%. The pattern he identified is the same one we see in student support: the gap is not in the quality of support available. It is in the accessibility of that support.

The AI provides an immediate first point of contact for students with questions about enrolment, course selection, deadlines, policies, support services, and academic procedures. Available at 11pm on a Sunday when the student is finally sitting down to sort out their enrolment issues.

This is not a chatbot providing generic answers. It is a system trained on the institution's specific policies, procedures, and support pathways. "What are the prerequisites for COMP201?" gets a precise answer, not a link to a catalogue.

Based on the student's enrolment data, academic history, and stated goals, the AI provides personalised recommendations: relevant courses, support services, study resources, and pathways. A first-year student struggling with statistics gets different guidance from a postgraduate student selecting thesis topics.

Josiah's insight here is critical. Personalisation in education is not about giving students what they want. It is about giving them what they need to succeed, framed in a way that they can act on. The AI needs to understand not just academic requirements but learning context: is this student working part-time? Are they a first-generation university student? Are they studying in their second language?

The AI monitors engagement signals (LMS activity, assignment submissions, attendance patterns) and identifies students who may be at risk of disengaging. This is not punitive surveillance. It is proactive outreach: "We noticed you haven't submitted the last two assignments. Here are your options and the support available."

Tim's adoption expertise shapes how these interventions are designed. An early warning that feels like monitoring drives students away. An early warning that feels like someone noticing and caring encourages engagement. The difference is in the framing, the tone, and the follow-up.

Universities have extensive support services: academic advising, counselling, financial support, disability services, career services, library support. Students often do not know what exists or how to access it. The AI acts as a navigator, connecting students to the right service based on their situation.

"I'm stressed about money" connects to financial hardship support, scholarship information, and budgeting resources. "I'm not sure this is the right course" connects to academic advising, course transfer processes, and career guidance. The AI does the routing that students often struggle to do themselves.

Counselling. AI does not provide mental health support, personal counselling, or crisis intervention. When a student indicates distress, the system immediately connects them to human support services. This boundary is absolute.

Academic assessment. AI does not grade work, assess learning, or make academic judgements. It supports the administrative and navigational aspects of student life, not the academic ones.

Replace human advisors. The goal is to handle the routine questions (80% of enquiries) so human advisors can focus on the complex situations that require expertise, empathy, and judgement.

Tim and I have learned that education AI adoption follows different patterns from enterprise AI adoption. Students are quick to try new tools but equally quick to abandon them if the experience is poor. Staff adoption is harder: academic staff worry about being replaced, and support staff worry about being deskilled.

The adoption strategy that works:

Students: make it frictionless. Integration with existing platforms (LMS, student portal, messaging apps). No separate login. No learning curve. Available where students already are.

Academic staff: make it invisible. The AI supports students in ways that reduce the administrative burden on academic staff without requiring them to change their teaching practice.

Support staff: make it empowering. The AI handles routine enquiries, freeing support staff for the complex, rewarding work. Position it as a capability amplifier, not a replacement signal.

Data privacy is paramount. Student data is sensitive. The system must comply with the Privacy Act, the Education and Training Act, and the institution's own data governance policies. Students must know how their data is used and have control over their engagement with the AI.

Cultural responsiveness matters. NZ tertiary institutions serve diverse student populations. The AI must be culturally responsive, not just technically accurate. This means understanding that a Māori student's support needs may include whanau engagement, that a Pacific student may value community-oriented guidance, and that an international student faces specific challenges around visas, employment, and cultural adjustment.

Continuous improvement is essential. Student needs change. Institutional policies change. Course structures change. The AI must be maintained as a living system, not deployed as a static tool.

Josiah's education technology experience and Tim's adoption expertise converge on this: the technology is the easy part. Building something students trust, staff support, and institutions can sustain is the real work. That is what we focus on.