On a warm lunchtime for most of the UK, colleagues from across the sector came together not just to compare notes, but to make sense of what AI is starting to mean in practice for the work professional services teams do every day. We touched on everything from the Jisc Professional Services curriculum design, governance, recruitment, transparency, and staff confidence. Although institutions are approaching these questions from different starting points, many of the themes are familiar: how to build confidence without forcing engagement, how to support experimentation without creating unnecessary risk, and how to keep professional judgement at the centre as new tools become part of everyday work.
Developing the AI in Professional Services curriculum
To kick off we discussed Jisc’s recent AI in Professional Services curriculum. I briefly reflected on the first iteration, which currently consists of three one-hour webinars covering core AI concepts, practical prompting and use, and wider issues such as ethics, copyright, data security, and environmental impact. Interest has been strong, with attendance in the high hundreds, and there was a real sense that the curriculum is meeting a genuine need across the sector.
At the same time, the discussion was refreshingly candid about how difficult it is to design learning for people who are at very different stages in their AI journey. For some, the material felt rich and timely; for others, especially in areas such as ethics, the pace could felt overly fast. The shared view was that this curriculum will need to keep evolving in step with participant feedback and with the pace of change in the AI landscape itself.
There was interest in developing more focused sessions aimed at particular professional services contexts, such as governance, finance, or student support, where the questions can be much more specific and practical. A recurring theme was the value of case studies and real-world examples: not simply to show what AI can do, but to help staff develop the judgement to decide when and how it should be used well. Colleagues are encouraged to share resources they have developed locally via the AI in Professional Services Jiscmail group, reinforcing the spirit of the community as a space where people can learn with one another rather than tackling the same challenges in isolation.
Managing approval, monitoring and the reality of ‘shadow AI’
Governance and approval processes formed another important part of the conversation. Colleagues shared how their institutions are trying to navigate the approval and monitoring of AI tools, often with IT and data protection teams playing a central role. What came through clearly was that many organisations are working hard to put sensible guardrails in place, especially where personal or sensitive data is concerned, but are doing so in a context where staff curiosity and experimentation are moving quickly.
Alongside formal policies, there was a lot of recognition that the real challenge often lies in communication. Rules may exist, but if they are difficult to interpret, or if licensing and third-party terms make the picture more complicated, staff can still be left unsure about what is actually permitted. That uncertainty can be hard to resolve in practice, particularly for busy teams trying to make sensible decisions in the flow of day-to-day work.
The discussion suggested that governance on its own is not enough. AI literacy, clear guidance, and a culture where people feel able to talk openly about how they are experimenting may do more to reduce risk than a policy document sitting unread on a shelf.
Creating engaging AI hackathons and safe spaces to learn
The discussion then shifted to the kinds of spaces that help people learn well. There was plenty of enthusiasm for small-scale hackathons and practical team exercises where colleagues can try things out together, compare responses from different large language models, and talk honestly about what works, what does not, and why. These activities were seen as especially useful because they make prompting feel less abstract and more rooted in real tasks and workplace scenarios.
A suggestion was that agent-based activities could be a particularly engaging next step, for example asking teams to build custom assistants around real day-to-day challenges: learning becomes more meaningful when it starts from familiar problems rather than generic examples, and it gives people a clearer sense of how AI might fit into their own context.
The overall emphasis was on creating safe and inclusive spaces for learning. The conversation was not only about energising the early adopters. It was also about making sure that colleagues who feel cautious, sceptical, or simply new to the topic have room to ask questions without judgement. Ideas such as peer groups and ‘learning triangles’ felt especially promising in that respect, offering a way to build confidence gradually and to make exploration feel supportive rather than performative.
Declaring AI use in professional services work
When should staff declare AI use, and what should that look like in professional services work? Some institutions are already encouraging simple declarations, similar to those used in academic contexts, that briefly explain which tool was used, why it was used, and what kind of support it provided. For some colleagues, this feels less about compliance and more about modelling an open and responsible culture.
What made this especially interesting was the recognition that the boundary around AI use is becoming blurrier all the time. As generative features are woven into the software people already use every day, it becomes harder to define where ordinary digital assistance ends and AI use begins. A drafting suggestion, a summary feature, or an embedded assistant may increasingly feel like part of the normal workflow rather than something separate. That raises important questions about whether current guidance will still make sense as these tools become more routine.
AI and the changing nature of recruitment
The final major theme was recruitment, where the conversation felt especially timely. Several colleagues spoke about the growing volume of AI-assisted or AI-generated applications and the practical challenge this creates for recruitment teams. Applications may appear polished and fluent but sometimes feel generic or only loosely connected to the role. That makes it harder to judge authenticity, fit, and effort, while also adding to the workload of those reviewing large numbers of submissions.
There was also an honest exchange about disclosure. Even where applicants are encouraged to be open about using AI, many may hesitate to do so if they fear it will be seen negatively. At the same time, the group recognised that AI can offer genuine support, particularly for candidates who find writing challenging or who benefit from help in structuring their ideas. The question, then, is not simply whether AI has been used, but whether the application still reflects the candidate’s own thinking, judgement, and care.
It should be noted that these are unlikely to be isolated questions for long. Across the sector, AI is beginning to prompt a wider rethink of recruitment processes as with academic assessment methods. If parts of application writing can increasingly be supported or automated, institutions may need to look again at which stages of recruitment genuinely reveal the qualities they care about, and how those processes can remain fair, proportionate, and meaningful.
Thanks to all those who attended and engaged so freely. We will be holding our last session before a short Summer break for the community meetings however the AI in Professional Services AI curriculum will be running across July and August. Details can be found on our Jisc training pages.
Our last community session of the year will take place on 24th June at 3:30 pm – 4:30 pm.
Find out more by visiting our Artificial Intelligence page to explore publications and resources, learn more about our communities and sign up for our AI Literacy training.
For regular updates from the team sign up to our mailing list.
Get in touch with the team directly at AI@jisc.ac.uk
