18 November 2025 - DIG25 Rad Bach
UCISA DIG25 Conference – Feedback
by Rad Bach
Attending the UCISA DIG25 Conference in Bristol was an engaging and valuable experience that gave me a clearer understanding of how the UK higher education sector is approaching digital transformation, and particularly the rapidly evolving landscape of AI governance. Although my initial expectations and first impressions were mixed, the overall event proved highly informative and a strong opportunity for professional reflection and networking.
Before arriving at DIG25, I was primarily looking forward to deepening my understanding of AI governance within a Microsoft tenant environment so I could feed this knowledge back into ongoing policy development within my organisation. With so many UK universities now exploring the boundaries of generative AI, I was particularly interested in practical guidance, governance blueprints, and structured frameworks that could be adapted or benchmarked against. I had anticipated a workshop-style environment with more hands-on examples or technical demonstrations, but instead the event leaned toward high-level knowledge sharing, organisational experience reports, and cross-institutional learning. Despite this, the content remained relevant and helped contextualise our own challenges within the broader sector.
My first impressions of the event were shaped by the venue choice, which initially felt slightly underwhelming. However, this was quickly offset by the relaxed atmosphere, approachable presenters, and a programme rich with interesting sessions. The informal tone of DIG25 actually made it easier to interact with both attendees and vendors, and it fostered a sense of open discussion rather than a rigid conference structure.
Several presentations stood out. Hearing from the University of South Wales, the University of the West of England, and others provided a clear view of how institutions are approaching cloud adoption, automation, and early AI experimentation. UWE’s discussion of Ivanti and Microsoft Copilot deployment was especially relevant to my own work. Their candid breakdown of Copilot’s limitations—not only around browser and webpage access but also its dependency on internal governance structures—helped set realistic expectations about its applicability and the maturity required for safe adoption. Similarly, their governance approach, involving small decision-making groups and rapid iteration, produced impressive engagement metrics: 4,000 interactions within seven months, with AI-assisted content reducing around 13% of service desk calls. This was both insightful and encouraging.
I also found the sector-wide theme of organisational silos and collaboration challenges relatable. Many institutions highlighted similar barriers between IT, academic teams, and governance bodies—making it clear that the human and organisational dimensions of AI adoption are as significant as the technical ones. Sessions from browser-based GPU VDI providers, VDI and cloud vendors, and AI platform integrators such as PAX Solutions added further perspective on the infrastructure and tooling side. Discussions about enterprise data layers, API maturity issues (including the complications surrounding EV connectors and vendor-specific APIs), and cost modelling helped reinforce that AI readiness is not purely a technical problem.
My top three takeaways from DIG25 are:
- Clear, high-value AI use cases are still difficult for universities to define, and in practice most institutions have not progressed far beyond student-facing chatbots and administrative automation pilots.
- Approaches to implementing language models vary widely, with each university using different vendors, governance models, and adoption strategies—reflecting how early the sector still is in its maturity curve.
- The financial equation remains uncertain, with many institutions weighing the cost of AI services and licensing against unproven or emerging value.
Finally, my main recommendation for someone attending DIG25 for the first time would be to focus on conversations as much as sessions. Speaking with academics, IT leaders, and vendors provides insight that cannot be captured in slides alone. The conference is also a social learning environment and approaching it with openness—rather than expecting rigid workshop structure—makes it more rewarding. Networking, comparing experiences, and discussing challenges informally are some of the strongest aspects of DIG25.
Overall, DIG25 was a meaningful and thought-provoking event that helped position my organisation’s AI strategy within the wider higher-education context and provided practical perspectives that I can bring back to my team.
