From AI use cases to implementation
Over the past two months, we've explored where AI is being used in information management and how to put those ideas into practice.
October's research brought together practitioner experiences, our community discussion and a hands-on experiment: building an agent to support subscription reviews.
Our report, "Building AI solutions for information work", explores the choices involved - from structured prompts to agents and applications - and what it takes to test and use them.
One useful reminder emerged: a tool can deliver value even if it helps just one information professional. Making it available across the organisation may require substantial extra work, so wider deployment is not always the right next step.
Building was quick. Checking took longer
As part of my research, I tried building an agent myself. Using fictional subscription and usage data, I set up an assistant to analyse the figures, flag issues and suggest next steps.
The first reports looked clear and well organised. But checking them revealed missing context, problems with conflicting figures and conclusions that needed another look.
Each change to the instructions meant another round of checks. Producing reports for different readers also meant making sure important details survived in the shorter summaries.
In the accompanying blog, "Building an AI agent was quick. Testing it took longer" I share what happened and why getting a convincing report was only the beginning. Check out my YouTube video to see it in action.
November: how do you keep up with sources?
Our attention now turns to the changing source landscape.
- How do you hear about new sources or changes to existing ones?
- Do you actively monitor developments, or investigate when a user brings a need to your attention?
- Which channels help - and could AI play a useful role?
Our next Community session, "How to monitor developments in the source landscape" (19th November 2026) will explore approaches, channels and tools, drawing on how information managers tackle this work in practice:
I look forward to hearing what works for you.
