evidoria

← Back to browse

Good practice Imported

AI Assistant Cuts First-Contact Resolution Time — University of Auckland's Institution-Wide Student Services Bot

New Zealand · Auckland · See the New Zealand profile · See the Auckland profile

Evidence: Descriptive / self-reported Top 81% 33/100 · Ask Evidence Copilot about this practice

Built with IBM Watson and university spinout ElementX, the AI Assistant handles 9,000+ conversations and 60,000+ searches monthly across the ~40,000-student University of Auckland, lifting first-time resolution 40% — though privacy and dependency concerns remain.

>9,000
Monthly conversations
>60,000
Monthly searches
>90%
Vendor-reported coverage/containment
40%
First-time resolution improvement
58%
Self-service to assisted-service ratio improvement
AI Assistant Cuts First-Contact Resolution Time — University of Auckland's Institution-Wide Student Services Bot

Details

Maturity
Established
Promoter
University of Auckland (Waipapa Taumata Rau), with ElementX and IBM Watson
Period
Staged rollout 2024–2025; formally announced October 2025
Keywords
higher education, student services, IT administration

Context

The University of Auckland sought to unify information from 90+ university digital domains into a single support channel across its five-campus, roughly 40,000-student institution.

Activities

The university partnered with its own Business School spinout ElementX and an IBM Watson-based retrieval-augmented-generation system to build the AI Assistant, staged through 2024-2025 and formally announced in October 2025.

Results

The assistant now logs 9,000+ conversations and 60,000+ searches per month, with vendor-reported coverage and containment both averaging above 90%. First-time resolution improved by 40% and the ratio of self-service to assisted service improved by 58%; the assistant now generates more interactions than the university's incoming phone lines. The project won a CAUDIT award and the Tertiary ICT Excellence Award for Technology Delivery.

Conclusions

The university's own privacy page discloses that conversations escalated to a human agent become identifiable records tied to a student's email or ID number. Separate student-feedback research found some students perceive chatbot answers on complex topics as generic, and raised concerns that dependency could undermine critical thinking.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years)
Staffing & skills
ElementX (University of Auckland Business School spinout), IBM Watson implementation team, University digital-services staff

Conditions for success

  • Unified information from 90+ separate university digital domains into a single retrieval-augmented-generation system rather than domain-by-domain chatbots
  • Staged rollout across 2024-2025 before formal institution-wide announcement

Common failure modes

  • Student-feedback research found some students perceive chatbot answers on complex topics as generic, and raised concerns about dependency undermining critical thinking
  • Escalated conversations become identifiable records tied to a student's email or ID number, a privacy consideration

Commonly funded by

Own resources / municipal budget

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful