Progress without quietly taking people backwards
Executive summary
IT Support is no longer judged mainly on how many tickets it closes or how quickly phones are answered. It is now judged on how it feels to get help. Was it easy? Was it clear? Did I feel understood? Did it actually solve my problem?
At the same time, organisations are under enormous financial pressure. Demand for digital services keeps rising, but budgets and headcount do not. In this environment, AI and automation are being positioned as the natural answer.
They can be, but they can also quietly make things worse.
The danger is not that AI and automation fail technically. The danger is that they succeed operationally while failing experientially. Users may be pushed away from people, into systems that are technically available but practically frustrating. Staff may be asked to rely on tools they do not fully trust or understand. What looks like progress on a dashboard can feel like regression in real life.
This paper takes a pro-choice position: AI and automation are powerful and necessary, but only if they are used in ways that respect how people experience IT support. It explores how IT support risks going backwards when technology runs ahead of human capability, and how keeping the customer journey and the voice of the customer front and centre is what separates progress from accidental harm.
IT support has entered the experience economy
For most of IT’s history, success was measured by output: tickets closed, calls answered, systems uptime, SLA performance. These are still important, but they are no longer what users judge most strongly.
People now judge IT by questions like:
- How easy was it to get help?
- Did I have to explain myself multiple times?
- Did it interrupt my work?
- Did I feel confident that someone was actually in control?
This shift is what many refer to as the experience-led economy. In this world, perception is as real as performance. You can meet every SLA and still lose trust if the journey feels clumsy or impersonal.
This is why Experience Level Agreements (XLAs) emerged: to measure things like effort, sentiment, and confidence alongside traditional metrics. Gartner has described XLAs as a way to connect service delivery to how it is genuinely experienced, not just how it is reported.
AI and automation now sit right in the middle of this shift. They increasingly shape the very first interaction users have with IT. That makes them powerful and risky.
Automation is not new – but what it now touches is
IT Support has been automating for decades:
- Password resets
- Account creation
- Software installs
- Device builds
- Monitoring and alerting
- Ticket routing
- Knowledge bases
None of this is radical. It is the backbone of modern IT service management, widely reflected in industry frameworks and standards.
What is new is that AI now sits at the front of the experience. Instead of quietly running workflows behind the scenes, it increasingly becomes the voice, the interface, and sometimes the decision-maker.
Generative AI can:
- Interpret what a user types
- Decide what they need
- Suggest solutions
- Draft responses
- Write knowledge
- Summarise cases
- Trigger actions
That changes the emotional contract between people and IT. Users are no longer just dealing with a system; they are interacting with something that sounds like it understands them.
This creates a dangerous illusion of competence.
Recent research and journalism have shown that AI systems can sound confident while being wrong, incomplete, or based on shaky sources. This is not a theoretical risk, it has been observed in live deployments.
When that happens in IT support, the damage is not just technical. It erodes trust.
How experience quietly goes backwards
Most experience failures in AI-driven support do not look like failures at first. They look like efficiency.
“Deflection” that does not actually help
When systems are designed to keep users away from humans, they can push people through loops of articles, bots, and forms without actually resolving the problem. The user leaves frustrated, often returning through another channel.
On paper, contact volumes go down. In reality, effort goes up.
The curse of repetition
One of the strongest drivers of dissatisfaction in support is having to repeat yourself. If AI and automation do not pass context cleanly to a human when escalation happens, the user experiences the system as fragmented and uncaring.
Even if the issue gets fixed, the journey feels broken.
Over-standardising human problems
Some IT requests are simple. Many are not. They involve workarounds, timing, pressure, risk, or partial failures. When these get forced into rigid automated paths, users feel trapped.
The system may be efficient, but it is not kind.
Capability gaps
Not all users are confident with digital self-service and even digital as a concept. Not all staff are confident supervising AI. When organisations push ahead anyway, the gap between what the system can do and what people can comfortably use widens.
That gap is where experience collapses.
What research tells us about AI and customer experience
Studies of generative AI in customer service show a set of paradoxes. AI can reduce costs and increase consistency, but it can also reduce perceived empathy, trust, and confidence when not handled carefully.
In other words: it is possible to improve service quality on paper while making customers feel worse.
This matters deeply in IT support, where users are often already under stress; locked out, blocked, late, or worried about losing work. In these moments, being understood is as important as being solved.
The IT support lifecycle still matters
Service management standards such as ISO/IEC 20000 are built around the idea of a lifecycle: design, transition, delivery, and improvement.
AI does not remove the need for this discipline. In fact, it makes it more important.
Without lifecycle thinking:
- Bots are trained on bad knowledge
- Automation accelerates broken processes
- AI scales inconsistency
- And errors are reproduced faster than humans can catch them
The result is not digital transformation, it is digital amplification of dysfunction.
Why “experience-led” must be the anchor
If AI and automation are introduced purely as cost-saving tools, experience will almost always lose.
If they are introduced as experience-shaping tools, the questions change:
- Does this reduce effort for the user?
- Does this increase confidence?
- Does this make it easier to get help when things go wrong?
- Does this support staff in doing better work?
XLAs and experience management thinking exist to keep these questions visible. They ensure that organisations do not mistake efficiency for value.
Progress without regression
AI and automation absolutely belong in IT support. Demand is too high and budgets too tight to pretend otherwise.
But progress in an experience-led economy is not about replacing humans with machines. It is about using machines to remove the friction that stops humans from being effective.
The real risk is not that AI will fail. The real risk is that it will succeed in the wrong way.
If organisations let AI become the gatekeeper, the judge, and the voice of IT without putting experience, trust, and human capability at the centre, they will quietly make support harder, colder, and more brittle, even while their dashboards look better than ever.
In an experience-led world, that is not progress.
Conclusion
AI and automation are not threats to IT support; unthinking deployment is. The experience-led economy does not reward the organisation that closes the most tickets, but the one that makes it easiest for people to get back to being productive with confidence and trust. When automation is used to remove friction, support becomes calmer, faster and more human. When it is used to deflect, gatekeep or hide complexity, it quietly transfers effort from the organisation to the customer. The same technology can therefore either elevate or erode experience depending on how it is framed. In a world of tightening budgets and rising demand, the real leadership challenge is not whether to use AI and automation, but whether we are disciplined enough to let experience, not efficiency, decide where they belong.
