PASS assistant
AI assistant streamlines support for student leaders
A practical innovation from UNE's PASS team demonstrates how thoughtful change management can turn AI fear into adoption success. The project tackles a common challenge across the university: managing high-volume administrative queries while empowering users to embrace new technology.
The challenge of supporting a student workforce
The PASS (Peer Assisted Study Sessions) program targets traditionally difficult undergraduate units, recruiting high-achieving students to facilitate weekly study sessions. With roughly 100 student leaders at any given time and a 30% turnover rate per trimester, the program faces a unique operational challenge.
"For many of them, it's their first job," explains Rochelle Dowsett, Data and AI Project Coordinator for PASS. "When we train them, we're not just having to teach them things like how to avoid reteaching or session management strategies. We also have to teach them things like how to fill out a timesheet, what SharePoint is or how do I make a Kahoot quiz."
The intensive two-day training block inevitably leads to forgotten details and a flood of follow-up questions. In the early weeks of each trimester, the PASS team was spending upwards of 20 hours per week responding to basic administrative queries from new PASS Leaders.
An AI assistant for on-demand answers
The solution was an AI assistant trained on all PASS documentation, policies, procedures, and workflows. Built to provide instant, 24/7 support, the assistant handles the repetitive questions that previously consumed significant staff time. Rochelle developed and trialled the assistant as part of her post-graduate studies then rolled it into Madgwick, once that became available to all staff.
"I trained it on all of our documentation, our policies, procedures, workflows, all the stuff that we would tell them," Rochelle says, "in hopes that it can tell them not just the first time, but the second time or the fifth time or the 10th time, and free us up to do more of the higher-level work."
The real work: overcoming fear and building trust
Technical accuracy wasn't the biggest hurdle - user adoption was. "Our biggest stumbling block with it hasn't been accuracy. It's been fear," Rochelle notes. "As students, they've been getting quite a lot of mixed messaging over the last few years from various places about whether or not they should even be looking at AI or touching it."
Using lean startup methodologies, Rochelle implemented targeted change management strategies, including beta testing and a "what it got wrong" log where leaders can flag inaccuracies.
PASS Project Specialist Francie Finn added a crucial modelling strategy. "When they come to us with a question which we think they could have gone to the bot with, we now ask the question to the bot and then send that back to the Leader saying, 'we've asked it this question and this is what it said, and we agree with this', just to give it validity and allow them to see the way it's used and why it's useful."
Early results and wider potential
Adoption has increased steadily, particularly since incorporating AI training into professional development. The team is already seeing a significant reduction in administrative queries, creating time for more strategic work and additional AI assistant development.
"I can really see how this kind of concept could be applied across a wide range of uses across the university," notes Shannon Tyrrell, AI Innovation Lead at LabNext70.
The PASS assistant demonstrates that successful AI implementation often depends less on technical sophistication and more on understanding user concerns and building trust through thoughtful change management.