DIGIT’s inaugural AI Business Summit brought together senior technology, data, and business leaders in Glasgow to examine how organisations can move beyond AI experimentation and turn emerging capability into measurable business value.

Across the day, speakers explored the practical realities of AI adoption: how organisations can build the right foundations, identify meaningful use cases, govern systems responsibly, and ensure new tools are embedded into the way decisions are made.

Across three packed sessions featuring 20+ expert speakers, there was a whole lot to unpack. Below, we’ll take a lot at some of the main themes that emerged through a day of shared learning, discussion, and high-level networking.

Beyond the hype cycle

A recurring theme throughout the summit was that AI adoption is no longer primarily a question of whether organisations have access to powerful tools. The harder challenge is whether they are ready to use them well.

In the opening breakfast briefing, Samer Sallam, Data Practice Lead at Accenture, explored the importance of AI fluency as a human and organisational capability. The session positioned effective AI use as something that depends on leadership, culture, judgement, and responsible practice, rather than prompt tips or tool onboarding alone.

That same challenge was reflected later in the morning by Martin Paice, Technology Practice Lead at Valcon UK, and Simon de Timary, Data Practice Lead at Valcon UK, who examined the gap between AI ambition and value realisation. Their session focused on the foundations organisations need before AI can deliver meaningful results, including high-value use cases, organisational readiness, platform readiness, data foundations, and responsible AI.

The message was clear: AI value does not emerge simply from launching pilots or adopting new platforms. It depends on clear business priorities, the right data, suitable technology foundations, and the organisational capacity to turn experimentation into delivery.

Turning insight into action

Several sessions also returned to the very pertinent issue of where AI value gets lost inside organisations. It’s a huge problem – a recent RingCentral report found that while 87% of UK organisations view AI positively, only 16% have fully deployed AI-powered digital workers, significantly lagging behind US businesses (41%). More than half (54%) of UK organisations remain in research, exploration, or pilot stages.

Shruti Sharma, Director of Data, Analytics and AI at Save the Children UK, argued that the operating model is often the bottleneck. Her session explored the disconnect between faster intelligence and slower organisational decision-making, highlighting how traditional reporting and approval cycles can prevent organisations from acting on insight at the speed AI makes possible.

This theme connected closely with the wider summit agenda, which framed AI implementation as a move from uncertainty and ambition towards impact and value. Taken together, the sessions suggested that the next phase of AI adoption will require organisations to rethink how decisions are made, not just how information is produced.

Building on real foundations

Another strong theme was the need to ground AI in real business problems and real operating environments.

Scott Ogilvie, Global Director of AI Strategy at Wood Group, explored this through the lens of asset and safety-critical industries, including oil and gas. His session focused on the need to apply AI to practical use cases where safety, sustainability, efficiency, and risk are central concerns.

Examples included maintenance optimisation, backlog optimisation, predictive maintenance, AI-supported design document review, and environmental monitoring. In each case, the emphasis was not on AI as a standalone innovation, but on how it can support better decision-making in complex environments where human oversight remains essential.

Recommended reading

• AI Agents Are Transforming Enterprise AI Adoption
• Cutting Corners on Context Costs AI Agent ROI, Gartner Warns
• Report: AI Agents Are Being Deployed Faster Than we Can Govern Them

That practical framing echoed the Valcon session’s emphasis on starting with high-value use cases and building only what is needed to support them. Rather than attempting broad transformation in one move, Martin Paice and Simon de Timary pointed towards focused, incremental adoption: identifying a meaningful problem, fixing the necessary data, building the required platform capability, and using that work to test governance and adoption in context.

Infrastructure for the next wave

The summit also looked at the technical foundations supporting the next generation of AI systems.

Alan Gray, Principal Engineer at NVIDIA, closed the day out with a session on the parallel computing powering breakthroughs in agentic AI. He explained how generative and agentic AI rely on large-scale parallel computing, from the vector operations underpinning AI models to the infrastructure needed to support increasingly complex workloads.

The session placed agentic AI within a broader progression from perception AI and generative AI towards more autonomous systems, noting that these shifts increase demands on compute, memory, networking, and supporting CPU-based environments.

This infrastructure perspective added an important layer to the day’s broader discussion. While many organisations are focused on use cases, readiness, and adoption, the underlying technology is also evolving rapidly, creating new possibilities for more capable and autonomous AI systems.

Across the summit, the central message was that AI adoption is entering a more demanding phase.

The early wave of experimentation has shown organisations what is possible. The next challenge is turning that possibility into impact: selecting the right problems, preparing the organisation, building on reliable data, governing systems responsibly, and ensuring AI changes how work is actually done.