Artificial Intelligence

Guest blog by Richard Farrell, CIO at Netcall, as part of techUK's Cloud and Digital Infrastructure Focus Series 2026.

Guest blog by Gary Sinclair, Principal Solution Architect at Shaping Cloud, as part of techUK's Cloud and Digital Infrastructure Focus Week.

Guest blog from Alexis Evangelou, Director of Product & Solutions Europe & Middle East at Iron Mountain, as part of techUK's Cloud and Digital Infrastructure Focus Week 2026.

Guest blog by Vish Nandlall, Founder at Vish Nandlall Consulting, as part of techUK's Cloud and Digital Infrastructure Focus Series 2026.

Guest blog by Tijs Broeke, Public Policy Director at HP Inc, as part of techUK's Cloud and Digital Infrastructure Focus Week.

Who will be able to supply cloud services to European public bodies? How will the EU help businesses put AI to work? And will simplifying the digital rulebook make it easier to operate across borders? The European Commission’s 2027 work programme, published on 4 November, gives us a clearer sense of when Brussels will tackle these questions.

Government's Autumn Budget put around £700m behind Ai, across security, Sovereign AI, and regional adoption. While we're part chuffed, our team is now laser focused on what this means, what comes next, and what delivery will look like. Read here to find out more.

By John Newton, UK Head of Public Sector, Iron Mountain

What did we learn from this year’s NCSC Annual Review 2026?

Agentic AI governance is critical as UK organisations scale autonomous systems. Explore the technical, organisational, and policy controls required.

As energy systems become more digital, automated and data-driven, trust in these digital systems is now as critical as trust in physical infrastructure. Artificial intelligence (AI), advanced analytics, and digital twins already influence how stakeholders within the sector plan, operate and invest today. The question is no longer whether these technologies will shape the future of energy systems; but whether they are safe, transparent and in the interests of consumers when deployed at scale.

In this guest insight, Nimitha U of VE3 argues that the key question for agentic AI is how much authority businesses give it, not just what tasks it can complete. She sets out practical controls on permissions, reversibility, audit trails and human intervention, with agentic commerce in retail as an example.

In this guest insight, Nadiia Hulchuk of SoloWay Technologies looks at how AI is moving from pilots to everyday business use. She sets out the common mistakes to avoid, from overtrusting polished demos to underestimating long-term costs.

Building effective human oversight into agentic AI Most of the agentic AI systems I've seen proposed over the last year come with a reassurance attached, which is that a person will stay in the loop. A human approves the payment the agent has set up, signs off the report it has drafted, or checks the decision before it reaches a customer. It's a sensible instinct, and for a lot of regulated work it isn't optional. Under the EU AI Act, anyone deploying a high-risk system will need to make sure the people overseeing it are competent, trained and able to act, and in the UK the Data (Use and Access) Act has widened the room for automated decisions, provided safeguards are in place, including a way for people to obtain human intervention. What gets much less attention, in my experience, is whether that oversight can still happen once agents are in place, and whether anyone could show afterwards that it did. A lot of the review in today's workflows happens almost by accident, at the point where one person passes work to another. The manager reads the draft before it goes out, the second team looks at the case when it lands with them, and the payment sits in a queue until someone signs it off. Nobody designed most of those moments as controls, but they catch a good deal. Agents are very good at joining up steps that used to be handled by different people, and when they do, those moments tend to go too, usually without anyone having decided they should. In some set-ups the person and the agent are working on the same piece of work together, so there isn't a separate point of review to speak of. techUK's industry brief on scaling agentic AI recommends building observability in from day one, and that's right. But it answers a different question. Observability tells you what the agent did, and it doesn't tell you whether a person examined it, whether they had what they needed to catch a problem, or whether they could have stopped it. A record that someone pressed approve, or accepted a suggested change, looks exactly the same whether they read the work carefully or barely glanced at it. A colleague and I recently proposed a framework to the European Commission on what organisations would need to record for human oversight of AI-supported work to be demonstrated, and what an inspector could reasonably conclude from those records. One of the things it sets out is that the phrase "a human reviews it" can hide several quite different situations. Sometimes the check exists in the procedure and nothing shows it was carried out. Sometimes it happens, but the reviewer isn't really in a position to spot a problem, because they don't have the context, the time, or the standing to send the work back. Sometimes it's done properly on some of the work and not the rest, and sometimes the person doing it isn't the one the procedure says should be. Each of those calls for a different fix, and a policy that says review takes place doesn't tell you which one you have. For organisations deploying agents, I think a few practical things follow from this. The first is to decide where a person needs to check the work before the workflow is automated, rather than discovering afterwards that the checks have gone. Mapping the process before deploying agents, which techUK's brief also recommends, is the natural point to do it, because it's when you can see where work currently changes hands and which of those moments are doing real work. The second is that where review has to be shown as a distinct step, the workflow needs two things built into it, somewhere in the process where the review can genuinely happen, and something that records whether it happened. Neither is much use without the other, and neither can be added cheaply once the system is live. The third is to record more than the approval. Whether the reviewer had the information and the time they needed, and what share of the work the check actually reached compared with what it should have, are what turn a log entry into evidence that oversight happened. And the fourth is to be honest about the steps that don't need a person at all. Some work is better handed over completely, and keeping a nominal approval on it costs time while adding very little. Taking the check away deliberately, and saying so, is a better position to be in than keeping one that nobody can show is doing anything, and that only adds to approval fatigue. This doesn't need new technology. It needs the oversight to be designed in alongside the agent, rather than assumed to survive it. I think the organisations that do this well will find they can give agents more responsibility over time, because they'll be able to show where people still matter in the work, and that the people were there. Bio: Dr Jack D. Saunders is a partner and Technical Director at Trident Intelligent Solutions, where he leads a team building AI and data systems for organisations in regulated sectors. He holds a PhD in machine learning, and his recent research looks at how human oversight of AI can be specified and evidenced.

In this guest insight, Tatjana Radisic of UST argues that as AI agents increasingly work together across processes, governing each agent on its own no longer holds up at scale. She calls for governance at the level of the service the agents deliver, drawing on the FCA’s operational resilience regime as a model.

In this guest insight, Barley Laing of Melissa looks at how the rise of sovereign and regional clouds is pushing organisations to silo customer data across jurisdictions, which damages data quality. He argues that global data quality platforms with fuzzy matching can keep siloed records accurate and consistent while staying compliant.

In this guest insight, Dr Ilya Burkov of Nebius argues that the UK has world-class science but needs the infrastructure to turn AI-driven discovery into companies and treatments. He sets out three priorities, compute, connected health data and faster translation, drawing on examples from London startups like Basecamp Research.

In this guest insight, Jiahao Sun of FLock.io explains how decision models, small AI models that give fast, structured answers rather than generating text, offer a cheaper alternative to large language models. He argues that their real value lies in running on local devices, keeping sensitive data private for sectors like government, finance and healthcare.

Nicholas Denning of Diegesis weighs up how organisations should modernise legacy IT when rising costs, unsupported systems and stretched teams collide with fast-changing AI tools. He argues for clear governance, sound architecture and investment in skills, warning that waiting for better tools risks never starting at all.

Today's Budget was the first opportunity for the new Andy Burnham government to set out their plans for the UK economy. Here's what the Budget means for tech

Rob Gittins of SecureCloud+ argues that AI-driven threats and recent attacks on government suppliers make cyber resilience an urgent priority. He calls for government and industry to build secure-by-design infrastructure and AI-ready defences together.

Matt Culpin of IEG Group argues that the most valuable public sector AI is built into existing workflows rather than offered as a standalone tool. He shows how it can support planning, healthcare and care assessments while keeping professionals in control.

Leanne Moorhouse of Pimloc looks at how rising subject access request volumes, particularly in social care, are putting pressure on public sector teams working to fixed statutory deadlines. She argues that AI-assisted redaction can take on the repetitive groundwork, while people stay responsible for every disclosure decision.

George Marlow of Telefónica Tech UK&I argues that government should use sovereign AI where control matters most, rather than by default. He explains how it fits within a wider enterprise AI platform, drawing on Telefónica Tech’s work with Cosine on Britain’s first sovereign frontier AI model.

Luke Craven of Simpson Associates argues that councils should start with the problem they want to solve, not the technology, when turning the AI Opportunities Action Plan into practical action. He sets out the data, governance and people foundations needed to move from experimentation to measurable outcomes.

How can industry support government to implement AI effectively? Helping to reshape procurement would be a great start.

Nimitha, AI Researcher at VE3, looks at why AI case recording tools often stall between pilot and everyday use in public services. She sets out what public bodies and suppliers need to get right from the start, from involving frontline teams to keeping practitioners in control of their records.

Scott Wylie of Whitespace draws on work with Surrey County Council, analysing thousands of contracts for local government reorganisation, to show what it takes to move AI from demonstration to operational use. He argues that clear scope, traceable findings and evidence-led decisions about scaling are key to building trust in public services.

Joe Baker of dxw argues that AI in government should start with the service, not the technology, because AI amplifies whatever practice is already there. Drawing on dxw’s public sector work, he shows why AI works best when it’s added to good practice, not used in place of it.

Aveva's Théo Farineau argues that industrial AI could underpin the Prime Minister’s ambition to “re-industrialise” the UK, but only if infrastructure and innovation clusters can share data as a connected system. He calls on government to work with industry on common standards, accountability and pilot funding to unlock trusted data sharing at scale.

Dr Shameena Bonomally of Page White Farrer explores how quantum, photonic and neuromorphic computing are emerging alongside silicon, each suited to different AI workloads. She argues that the UK should plan for how these technologies integrate from the outset, putting it in a position to lead on hybrid systems for AI.

Elis Thomas, from techUK, argues that AI convergence with other frontier technologies happens on the chip, which makes semiconductors central to the UK’s AI ambitions. He sets out the UK’s strengths and the short window to use them, and calls for a joined-up convergence strategy from government.

Welcome to the latest edition of Dispatch from Brussels, techUK's bi-monthly update on the European policy developments shaping the operating environment for UK technology companies.