Document Processing
University of Auckland researchers were spending too much time navigating grant, ethics, and export-control requirements because funding applications and compliance reviews involved multiple forms, committees, and rules that delayed research work.
Reported impact
The hackathon proved that agentic AI can reduce research-administration effort, with high routing accuracy and error detection, and it led directly to continued IBM-university work on production-grade tools for researchers.
The University of Auckland worked with IBM to frame research administration as an agentic AI problem, then used a hackathon with Master of AI students to prototype solutions on IBM watsonx.ai and IBM watsonx Assistant. IBM and the university supplied use cases from the Research and Innovation Office, along with evaluation criteria, mentorship, and technical support so student teams could build agents that review grant applications, ethics applications, and export permit submissions. The prototypes use publicly available guidance and few-shot prompting to produce context-aware feedback aligned to reviewer expectations, helping route applications to the right committee and detect known errors before human review. After the hackathon, the university and IBM continued collaborating on enterprise versions of the solution hosted on IBM Cloud, turning the event from a one-off exercise into a low-cost, scalable research-administration platform for ongoing use.
Research application materials and guidance documents are fed into watsonx.ai-powered agents, which use few-shot prompting to review content, flag known issues, and route submissions to the correct committee; the resulting feedback and workflows are then served through enterprise solutions hosted on IBM Cloud.
“By employing data from publicly available documents and using few-shot prompting to align with reviewer expectations”
“Hosted on IBM Cloud, the solutions using IBM watsonx operate at a low cost, which makes them scalable and on-demand tools for the University.”
“The proof-of-concept solutions developed by the students demonstrated the value and feasibility of agentic AI solutions for research administration”
“Using IBM watsonx.ai, competing student teams set out to create innovative, AI-powered agents to streamline the application and review processes for grant applications, ethics applications and export control permits.”
“the AI agent routed 95% of them to the correct committee and detected 85% of known errors”
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
IBM Sample Jan 2, 2026 National Agency for Education, Lithuania Lithuania’s National Agency for Education needed to extract, standardize, and compare information from thousands of mixed-format school management reports because manual extraction took hours per report and made… 14 key criteria tracked by AI solution IBM Cognos Analytics · IBM watsonx.ai 
Microsoft Sample Oct 15, 2026 Bar-Ilan University Students who miss lectures struggle to reconnect new lessons with prior material when course information is spread across notes, recordings, slides, and readings, and lecturers need a reliable way to answer routine… 71 courses Azure AI Services · Microsoft Azure · Microsoft Copilot Studio IBM Sample Sep 24, 2026 StarLink StarLink needed to eliminate manual effort in procurement processes such as email intake, purchase-order validation, document extraction, and quote comparison without losing governance or visibility across enterprise… 11,500 hours projected annual time savings across purchase order and quote processes Automation Anywhere · Docling · IBM Bob 