What’s generally useful in day-to-day work, and where are the boundaries?
The first AI in Practice session in March brought LIANZA members together to share how they are using AI in their day-to-day work and where they draw the line.
This session featured practical insights from Lynda Drumm (PwC) and Debbie Ewen (Waitaki District Libraries), and was facilitated by Erin Cairney, Chair of the LIANZA AI Committee.
This wasn’t a traditional webinar. It was a practical, member-led conversation focused on real examples, what’s working, and what isn’t.
WHAT AI IS BEING USED FOR
Across public, academic, and special libraries, AI is already being used in meaningful ways. The strongest theme was using AI to get started, save time, and generate ideas, rather than relying on it for final outputs.
Some of the practical use cases shared included:
Drafting emails, reports, and training materials
Summarising newsletters and multiple sources into quick briefings
Fixing Excel formulas and troubleshooting code
Generating SQL queries and improving reporting workflows
Supporting customers as a digital learning tool
Assisting with cataloguing decisions (as a starting point)
Creating marketing content and event materials
Analysing large datasets to identify trends and insights
Using project-style workflows to organise documents and notes
These are all areas where members are finding immediate, practical value.
WHERE AI IS NOT WORKING (YET)
Members were also clear about current limitations.
AI can produce results that look convincing but are incorrect or incomplete, particularly in research contexts. Some tools still struggle with complex or domain-specific tasks, and outputs often need to be checked against trusted library resources.
This reinforces a key point from the session: AI can support the work, but it does not replace professional judgement.
WHERE ARE BOUNDARIES BEING DRAWN?
There was strong alignment across the group on when not to use AI:
When working with sensitive or confidential information
When accuracy is critical
When human or professional judgement is essential
When there are concerns about how data is being used
Members also raised important considerations around:
Māori data sovereignty
Bias and fairness
Data provenance and trust
PRACTICAL TIPS TO TRY
A few simple approaches stood out:
Start with low-risk tasks and build con fidence
Treat outputs as a draft to refine, not a final answer
Ask follow-up questions and challenge the response
Cross-check against trusted sources
Try different tools for different tasks
WHAT’S NEXT?
At the end of the session:
60% said they want to try something new
34% plan to talk to others in their team
33% want to refine how they are already using AI
This reflects a sector that is actively exploring but doing so thoughtfully.
LOOKING AHEAD
The prompts used in this session are available here, so you can try these approaches yourself.
The next session in the AI in Practice series will explore LIANZA’s professional values in an AI-shaped information environment, building on the themes raised in this discussion.
