The UK's biggest Service Desk and ITSM show is done for another year. As ever, this year's SITS event at London Excel was buzzing- with a definite increase in attendees. If you couldn't make it (or even if you did) here's a recap of some of the things that were happening and the topics discussed.
This is probably not a big surprise for anyone. As expected, most vendors featured agentic AI headers on their stands. Also, a good chunk of the seminar timetable was focused on AI (and I’ll bet that many other sessions mentioned AI). On the face of it, nothing much has changed from last year. AI messaging is front-and-centre.
But the conversations have shifted. Last year was more about “What is AI?”, “Should we be using AI?” and “How should we be using it?”
Now, people are past the hype and asking more pragmatic questions: “How do we make it work?”, “How do we make it safe?” and “How do we scale it up?”
The shift is from hype to real-world results. To get the results, people are focusing attention on how it fits into their service management ecosystem. They want to know how they can operationalise AI safely. They want to know what the implementation path looks like. And they want to know about AI governance. Questions about guardrails are becoming almost as common as questions about what the technology can do. And rightly so.
Executive intent to adopt AI is high—they can see the benefits of automating large volumes of work that was previously out of reach of traditional workflow automations. Across organisations in the market, trust in AI is steadily increasing.
However, objective readiness (as opposed to perceived readiness) is still low. When agentic AI is layered on top of fragmented processes and inconsistent data, it fails to deliver reliable outcomes. It amplifies existing systemic issues. These issues need to be address before applying agentic AI safely.
If you want to dig into your own organisation’s AI-readiness, check out this AI Readiness Assessment. Across 30 multiple choice questions, it quickly assesses the foundational capabilities that are needed for a successful AI implementation. It takes about 5 minutes to complete and at the end you’ll get a short report that pinpoints your organisation’s strengths and weaknesses. And suggests clear steps to improve readiness.
Data readiness is the #1 critical success factor impacting AI implementations. Garbage-in-garbage-out (GIGO) has always been true in the IT industry. With AI, it’s more true than ever—because AI multiplies the scale and consequences of bad data. Part of the danger lies in the fact that an AI that leverages bad data will confidently produce bad answers and take bad actions (instead of saying “I don’t know”).
So good data really matters. Data of sufficient volume and accuracy (although accuracy is the more important factor) is the foundation upon which trustworthy, effective, and governable AI outcomes are built.
But what I saw at SITS was that many vendors are skipping over the data readiness issue to focus on AI features. “Plug-and-play AI”. “Out-of-the-box AI features”. “So long as your data is good…” These are dangerous oversimplifications.
The reality is that only the most basic (and generic) AI-powered functionality can be delivered using a pre-trained vendor AI model alone. To drive the most powerful and transformative AI use cases, AI needs access to your own operational data (incident tickets, changes, assets, end user information, and knowledge) to understand operational context, make informed decisions and execute actions safely. And that data needs to be right.
At Hornbill, it’s an important issue that we’ve been working to solve for our customers. Our Flight Check service is a unique, AI-powered data analysis service designed to ensure your organization’s data is fully agentic ready. Delivered by Hornbill Professional Services and leveraging HAi Labs' proprietary Machine Learning models, this service provides the essential foundation for a successful AI implementation.
In his session on Managing AI Risk and Uncertainty in Service Delivery, Valence Howden of Infotech Research Group made a very important point: Applying AI governance after AI implementation is too late.
It’s too late because agentic AI works fast, and at scale—accessing service records, employee data, workflows, knowledge bases, IT systems and business systems. When things go wrong, they go wrong fast. Without AI governance designed-in from the start, organisations risk:
Yet, there were few other mentions of AI governance at the event. Plenty of talk about AI ethics and what can go wrong. Not so much about how to prevent it. AI Governance is the key to preventing the AI “scare stories” that we continually see in the press.
To get AI governance right, organisations must treat governance as a foundational part of AI adoption—not an afterthought. Hornbill’s approach embeds AI governance directly into our AI architecture, workflows and operational processes from day one. This includes defining clear guardrails for what AI can access, what actions it can take, and where human approval is needed before actions are executed. Human-in-the-loop controls are particularly important for high-risk actions relating to business critical systems.
Effective AI governance also depends on strong control over permissions, auditability and operational visibility. Organisations need role-based access controls, runtime oversight, escalation paths, policy enforcement, and detailed reporting to ensure AI remains secure, explainable and accountable. And let's not forget the ability to report on the ROI of AI.
One of the highlights for us was the presentation from one of your customers: Warwickshire County Council. In the session, Emily and Alex from WCC explored how they’re transforming service delivery with a different way of thinking placed at the heart of their transformation journey: systems thinking.
Systems thinking is an approach to problem solving that views organisations, processes and teams as interconnected parts of a wider system rather than isolated silos. It focuses on understanding relationships, dependencies and feedback loops to improve how the whole organisation operates, enabling more effective, efficient and joined-up service delivery outcomes.
For Warwickshire County Council, this meant breaking down disconnected workflows and creating more joined-up, end-to-end service delivery across teams. By redesigning processes holistically and automating repetitive cross-functional activities through Hornbill, the council improved collaboration, reduced inefficiencies, accelerated service delivery and created a stronger operational foundation for their future improvement and transformation plans.
A quick survey of the audience showed that only a handful of people in the session were already familiar with systems thinking. It’s not a new idea, dating back to the 1950s, and it has gained a foothold in the public sector, particularly in citizen-facing services. It’s brilliant to see the systems thinking philosophy applied to employee services within a council, and we’re hoping that more councils will recognise the value of this perspective when approaching their own transformation projects.
If you’d like to know more about it, I’d recommend this keynote speech from John Seddon of The Vanguard Method, who is a provocative advocate of the systems thinking philosophy.
Read the Warwickshire County Council case study here
Almost absent for the last couple of years, I noticed an uptick in mentions of ITIL this year, which is no surprise as the launch of ITIL 5 was announced back in February. It’ll be the biggest overhaul of ITIL since 2019—and plenty has changed since then. As the rollout is phased, the main bulk of the new content is yet to be published, but it will reflect the rapidly changing landscape of service management:
Being a phased roll-out, we can expect a rising wave of renewed interest in ITIL over the remaining months of 2026.
If you were at SITS and have any observations you’d like to add, comment below. If you didn't have time to talk to us at the event, please don't hesitate to get in touch at hello@hornbill.com.