The Jisc AI in Professional Services Community meeting, held on 17th December 2025, brought together colleagues from across the sector to reflect on where we are now with AI adoption – and where we need to go next. With AI tools becoming increasingly visible in everyday workflows, the discussion focused less on hype and more on implementation: how to deploy AI safely, ethically, and at scale, while genuinely supporting staff.
Across the conversation, participants shared practical experiences, challenge with governance and tool selection, and emerging approaches to building confidence and capability. A consistent message emerged: success with AI depends as much on culture, clarity and collaboration as it does on technology.
Some key insights:
Automating the drudge work, not the human work
A recurring theme was the use of AI to remove repetitive administrative tasks, freeing staff to focus on work that benefits most from human judgement and empathy. One member shared early experiences with Google Workspace Studio, a new (and still evolving) automation tool that can read incoming emails, draft responses, update spreadsheets, and pull information from multiple sources automatically.
While the potential productivity gains were clear, they were candid about the current limitations. The tool remains “buggy”, reinforcing the importance of careful piloting rather than rushed deployment. Automation, the group agreed, should be introduced incrementally and with close monitoring, ensuring it genuinely supports staff rather than introducing new risks or frustrations.
Security, trust, and clear boundaries
Security and data protection underpinned much of the discussion. Participants stressed that AI adoption in professional services must start with enterprise-grade safeguards, including encryption, clear data handling policies, and assurances that prompts and outputs are not used to train external models.
A member outlined a comprehensive approach combining technical controls with staff guidance and training. Policies covering privacy, copyright, and intellectual property have been updated, while AI literacy has been embedded into CPD sessions and internal communications. A clear and simple message sits at the heart of this approach: Gemini (their institutional AI tool) is the only approved chatbot for sensitive or secure use, and staff should not input confidential information into other tools such as ChatGPT or Claude.
This clarity was widely welcomed as a way to reduce uncertainty and build confidence among staff navigating a rapidly changing AI landscape.
Secure platforms and local control: The role of nebulaONE
Another member shared the University of Manchester’s experience implementing nebulaONE, a secure AI platform hosted within the University’s own Azure environment. By providing access to multiple large language models without exposing data externally, nebulaONE enables experimentation while maintaining institutional control.
An added benefit is the ability to create AI agents within this secure environment, opening up possibilities beyond individual productivity tools. Participants noted that platforms like nebulaONE offer a compelling model for institutions seeking to balance innovation with robust governance.
Choosing tools wisely in a noisy market
Several participants highlighted the growing challenge of tool selection. One member described pressures from senior stakeholders to adopt well-known or popular AI tools quickly – sometimes without sufficient evaluation. While understandable, this urgency can obscure whether a tool actually meets organisational needs or performs as expected in practice.
This member emphasised that some tools appear impressive on the surface but fall short when examined closely, making due diligence essential. Another member then shared the concept of discovery groups: cross functional teams that systematically review and filter requests for new AI tools before adoption. this approach resonated strongly as a way to manage demand, set expectations, and reduce the risk of costly missteps.
Beyond IT: Use cases, change management, and capacity
Participants reflected on the limits of technology-led rollouts, noting that simply “turning on” AI tools rarely delivers transformation on its own. Many potential benefits sit within broader continuous improvement work, requiring time and space for staff to articulate pain points and explore how AI might help.
Finding that space is often difficult, particularly for teams already under pressure. One member noted that meaningful use cases, rather than abstract promises of efficiency, are critical for winning hearts and mind, especially amid understandable concerns about job security and change. Positive examples are emerging, particularly in learning design and online education, but these successes rely on close collaboration rather than top-down deployment.
Piloting with purpose in a volatile landscape
The group acknowledged the volatility of the AI vendor market. With major providers such as Microsoft and Google rapidly integrating new features into existing enterprise platforms, there is a real risk of investing in third-party tools that soon become redundant.
As a result, many institutions are favouring pilots, trials, and short-term experimentation over long-term commitments. This cautious approach gives teams space to explore what works while keeping future options open, particularly when provider roadmaps suggest similar features may soon be available as part of standard enterprise tools.
Looking ahead: Collaboration and shared learning
The session closed with a strong sense of collective momentum. Planned next steps include piloting AI tools in professional services contexts, gathering and sharing real-world use cases, and continuing to develop tailored ethical guidance for staff and students.
Perhaps most importantly, the discussion reinforced the value of cross-institution collaboration. By sharing lessons learned, both successes and setbacks, the community can move more confidently, and more responsibly than any single organisation could alone.
As AI continues to evolve, the focus is shifting from whether to adopt these tools to how to embed them thoughtfully. With clear governance, purposeful experimentation, and open dialogue, professional services teams are well placed to shape AI in ways that genuinely enhance their work – and the communities they serve.
This was our last session of 2025, and we will pick up in the new year after everyone has had a well deserved break! Our sessions for next semester will be at 3:30-4:30pm on:
- 14th January 2026
- 11th February 2026
- 11th March 2026
We look forward to seeing you all in the new year!
Useful links
AI in Professional Services: Purpose first, product second – Artificial intelligence – Why organisational need should drive AI adoption decisions
nebulaONE® – Secure, Affordable, Microsoft-based AI Gateway – A secure, institution-hosted platform for accessing multiple AI models
AI tools licensing review guidance – Practical guidance on copyright, licensing, and responsible AI use in research
AI Procurement Due Diligence – Artificial intelligence – Key considerations for assessing and purchasing AI tools responsibly
We did the math on AI’s energy footprint. Here’s the story you haven’t heard. | MIT Technology Review – An overview of how large-scale AI affects energy use and sustainability
Artificial intelligence and the environment: Looking ahead – Artificial intelligence – Exploring the future environmental implications of AI adoption
Artificial intelligence and the environment: Putting the numbers into perspective – Artificial intelligence – Contextualising AI’s energy use and environmental footprint
Widely used but barely trusted: understanding student perceptions on the use of generative AI in higher education – Research insights into ethical considerations for AI use across HE contexts
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