Why context is the foundation for AI success in service management

Great service management has always been human.

The incident no knowledge article covers. The request that doesn't fit any category and needs someone to work out what's really going on. The payroll administrator who's suddenly locked out of the system on payday.

These are the moments that matter in service management. They've always needed a person on the other end of the line. They're why we love what we do.

In day-to-day reality, however, time for these moments is scarce. Instead, teams answer the same ticket again and again: end users asking for information that's already in the knowledge base. Long backlogs form, and the operators who are there to solve problems end up becoming the bottleneck instead.

No one is happy with this status quo. End users get frustrated by long wait times. And service management professionals want more time to get ahead of the work instead of falling behind it – and more time for the challenges that need their expertise, not more password resets...

This leads us to, at its core, the greatest value AI can bring to service management. If we can hand the routine, everyday work to AI, we give operators their time back for the moments that matter: the work that demands the judgment, empathy, and care only humans can provide.

So far though, in our industry and beyond, AI’s promise hasn't always matched reality.

According to Gartner, only 48% of AI projects reach the production phase. And more than 80% fail – twice the rate of IT projects that don’t involve AI.

The reality is that most AI projects don’t survive contact with, well, reality. We can blame this on the technology, yet we’re seeing it become more capable every day. The real reason why AI projects fail comes down to the knowledge the AI stands on.

Context is everything

Ask your AI assistant how to set an out-of-office notification in Outlook and you'll instantly receive the answer you’re looking for. Ask it what you need to get a laptop ready for a new colleague starting on Monday, and while you’ll certainly get an answer, it probably won’t be anywhere near as useful.

The difference is context. The out-of-office query has a public answer: Microsoft has documented it and the AI has learned it. Answering the new starter query, on the other hand, requires knowledge only your organization knows: your onboarding workflow, who has to approve the laptop request, the models available, how long it usually takes to set up a new machine. Without that knowledge, even the most capable AI can only give a vague response.

One of the great truisms of AI is that you get out what you put in. If you want to win with AI, and free up time for the moments that matter, you need to build it on a foundation based on your organization’s unique context.

So where do you start?

The good news is you already have your foundation’s building blocks. Every incident you’ve logged, every workflow you’ve built, every knowledge article you’ve written has built the context AI needs. This information doesn't have to be the perfect AI foundation. It just has to be usable.

An AI you trust first needs a foundation you trust – one that's uniquely yours. Trust is what everything else hinges on, and it matters more than ever as AI in service management shifts from a tool that works with you to an agent that works for you.

Let go of the wheel, stay in the driver's seat

AI’s role in service management is already transformational. It drafts replies, summarizes long ticket trails, and helps your team get more done. But the way the work happens hasn't changed. Your people still run every step.

The bigger shift, and the one we're now on the cusp of, is AI that doesn't just help with the work but carries it out from start to finish.

Service management is entering the era of agentic AI. Operators will soon work alongside AI agents that don't wait to be asked – they take routine work off your plate and handle entire workflows.

This shift is a lot like handing the wheel to a self-driving car – both in what the technology can do and in the role you still play in getting it where it needs to go. Behind the wheel of a self-driving car, you don't let go and close your eyes until you've seen it safely drive the same road many times before. Trust is gained step by step.

The same is true for agentic AI – and your organisation's foundation becomes even more important. The AI has to know what has worked in your service organization, and what not. Your processes, your exceptions, and the solutions your team has repeated a thousand times.

An assistant that lacks context gives a poor answer; an agent that lacks context takes the wrong action. But with the foundation in place, every time you watch it get the route right, it becomes easier to let go of the wheel. Even then, you’ll always remain in the driver’s seat. Because while you're handing over the work, you're not giving up control.

The service teams that get the most from agentic AI will be the ones that pair context with control. That means AI that acts only within the rules you set, with every action visible, logged, and open to review. Where your foundation has gaps, your guardrails keep the agent on course. And when it gets something wrong, you can see exactly what it decided and why.

Back to the moments that matter

Taking your organization into the era of AI-powered service management, and aiming to be ahead of the curve, is daunting. There are new technologies to evaluate, new risks to manage, and plenty of unknowns.

But many of those unknowns become far more manageable once a foundation you can trust is in place.

Assistive AI can only suggest answers as good as the knowledge it draws on. Agentic AI goes further and acts on that knowledge independently. If your articles are outdated, incomplete or contradictory, AI won't fix that. It will amplify it, giving wrong answers faster and at greater scale.

So, don’t start with the technology, start with the knowledge it stands on. Review what's in your knowledge base, fill the gaps behind your most common tickets, and retire what's no longer accurate. Get that right, and every AI capability you add afterwards has something solid to build on.

For too long, the moments that matter have been the work that gets squeezed out. With the right foundation, and clear controls, AI can change this for the better – winning back time for the work that needs people: making judgement calls, fixing root causes, and moving from firefighting to prevention.

Great service has always been human. It always will be.

Service management now has a huge opportunity to win back the time that human service needs.

Want more Unlocked content?

Go to the Unlocked content hub. Watch the keynotes, read up on break-out sessions, and discover insights curated just for service management teams.