Skip to content
Start Free

Smarter Support, Same High Standards – XetNET Case Study

Joanna ByjośNov 12th, 2026

Let’s finally admit it loud and clear: the work that drains a team is often not the technically difficult work at all, but the endless return of questions everybody already knows how to answer. The password reset, the familiar configuration issue, the customer asking for the same explanation three other people needed that morning. You wouldn’t mark any of them as complex, yet together those tickets eat an enormous part of the day while the people with the most experience spend their time proving, once again, that they know the answer.

XetNET Case Study - Erno Räsänen - KonradoAI

That was part of the reality Erno Räsänen, Founder & CEO of XetNET, wanted to change. The company operates under its Finnish XetNET brand, providing web hosting, virtual servers, cloud, and capacity services. In other words, it runs exactly the kind of business where customers eventually find impressively creative ways to report the same technical challenges again and again.

But the problem was not simply ticket volume. It was repetition, inconsistency between people answering the same questions, and the very human fact that doing the same thing for the hundredth time is not exactly a recipe for an energized support team. What makes Erno’s experience with Konrado.AI worth looking at closely is that he did not approach AI as a shortcut to removing people from the process. He approached it as a tool that had to earn the right to work beside them.

The Expensive Part Was Not the Difficult Tickets

Hosting support naturally comes with difficult cases, and nobody expects those to disappear. The stranger problem is everything around them. When the same routine questions keep coming back, a skilled employee still has to open the ticket, understand what the customer means, choose the right wording, and send an answer. Multiply that by a large enough queue and the company can end up spending a heavy amount of expert attention on work that stopped being intellectually demanding a long time ago.

XetNET had another complication on top of that. Repetition did not always produce repetition in the answers, because different people naturally explained the same issue in different ways.

“Different people were providing different types of answers, even to the same repeated questions. Also the fact is that most of the questions customers kept asking were of the same kind which in one sense was a bit boring for us to answer.

“Boring” is not a metric most support dashboards bother tracking, but perhaps they should. Repetitive work has a cost even when every ticket is answered correctly, because it uses time without really using the skill of the person doing it. It can also make consistency harder than it needs to be, since every agent brings their own habits to questions that are fundamentally identical.

Konrado.AI gave XetNET a way to attack both problems at once. Instead of asking people to keep rewriting familiar answers forever, the team could let AI absorb a large share of that repetitive layer while still keeping the human side of support where it actually adds something.

Hiring More People Was Not the Only Serious Option

There is a very old reflex in support operations: when the queue grows, add another person to the queue. Sometimes that is exactly what a company should do. The trouble begins when hiring becomes the automatic answer to work that is growing mainly because the same questions are being repeated at a larger scale. At that point, the company is not necessarily buying more expertise. It may simply be buying more human hours for work that software could already help carry.

For Erno, that made Konrado.AI easy to justify as an experiment. Compared with the investment required to expand the support team, the proposition was – to use his words – extremely tempting, but the part of his answer that says the most about XetNET’s approach has very little to do with cost.

Think about a lumberjack:

a chainsaw doesn’t make a lumberjack obsolete.

It’s a tool which makes him or her more productive.

That exact same logic applies to support. Expertise is valuable when it is used to understand customers, solve unusual problems, make judgment calls, and deal with the cases that do not fit neatly into a pattern. There is nothing noble about forcing good people to manually reproduce the same answer all day when a tool can take care of much of that repetition and let them work at a higher level.

The Number That Matters Most Did Not Move

AI support is full of seductive metrics. Faster replies look good, ticket reduction looks even better, and everyone loves a graph that drops sharply to the right. Erno’s most interesting result may be slightly less obvious as a success metric, yet considerably more important in practice.

XetNET confesses to using about 60–70% of ticket answers made by KonradoAI, while its support feedback scores have stayed at the same level as before.

Customer support ratings after introducing Konrado AI?

UNCHANGED.

And that’s exactly the point.

There is a tendency to treat “no change” as a boring outcome, but a flat line can tell two very different stories. It can mean either that you failed to improve or that you refused to give up something that was already exceptionally good. The starting point decides which one you are looking at, and Erno’s own support rating leaves very little room for doubt.

“I’m ”the worst” support operator with an average rating of 9.72, so you’ll understand that there’s not much room for improvement. However, my biggest fear was that AI answers would have a negative effect.

With numbers like that, KonradoAI was never walking into a support operation that needed rescuing. The real test was whether it could take a meaningful share of the work without giving customers a single good reason to rate the experience any lower.

Trust Arrived One Ticket at a Time

Konrado.AI did not receive automatic trust simply because it had been deployed. It had to earn it, and apparently it did so in the most convincing way available to any support tool: by being right when an experienced human was fairly sure it was wrong.

XetNET Case Study - Quote - KonradoAI

That is probably a healthier description of AI adoption than blind enthusiasm could ever give us. If Konrado.AI can prepare a correct reply while Erno is still reading the ticket, the human process no longer starts from a blank page. The operator can review, adjust, and decide rather than reconstructing every familiar answer from scratch, which means AI becomes part of the thinking process instead of a separate machine waiting somewhere at the end of it.

The Real Test Comes When Everything Breaks at Once

A normal day is not always where support automation proves its worth. The real test comes when a software glitch affects many customers at once and the queue fills with different reports of the same problem. Even a team that knows exactly what happened can still be crushed by the simple arithmetic of having to repeat the answer ticket after ticket.

This is where Konrado.AI does more than speed up individual replies. It absorbs the repetition, leaving people free to focus on the incident itself, unusual cases, and customers who genuinely need more than the standard explanation.

“It’s an excellent tool. It allows us to handle tickets faster even with the tone we prefer. It processes repetitive tickets automatically – even on days when some software glitch has caused problems for a large number of customers. It helps our staff to stay motivated.

The sentence about motivation may actually be one of the most useful outcomes in the whole conversation. Support teams are often expected to absorb repetition as though boredom were simply part of the uniform. In reality, removing work that adds almost no intellectual value can have a very practical effect on how people approach the rest of the queue, especially when the same tool also helps them preserve the tone they want customers to hear.

That is where the lumberjack comparison comes back into focus. A better tool does not make the person less important. It makes it harder to justify wasting that person’s energy on the slowest possible version of the job.

Good Support Does Not Need More Heroics

For years, support operations have rewarded heroics. Someone stays late, someone clears an ugly queue, someone answers the same incident for the fiftieth time without losing patience, and everybody agrees that the team handled the pressure well. There is nothing wrong with resilience, but a mature operation should also be allowed to ask whether some of that pressure needed to exist at all.

XetNET’s experience with Konrado.AI points to a more useful version of efficiency:

  • The company can use a substantial share of AI-prepared ticket answers.
  • Customer feedback has held its ground.
  • Repetitive questions can be processed faster.
  • The preferred tone can be preserved.
  • The system remains useful even when a broader software problem causes many customers to ask for help at once.

That is the real promise of Konrado.AI for hosting companies. Not fewer humans for the sake of fewer humans, and not automation as a badge of modernity, but a support operation in which skilled people can spend more of the day doing work worthy of their skill while the repetitive layer is handled with far less friction.

We owe a particularly warm thank you to you, Erno, for letting us look over your shoulder as Konrado.AI found its place in XetNET’s workflows. Your skepticism, humor, and a practical way of looking at things gave us a much clearer sense of what it takes for a tool to earn real trust when usefulness has to survive contact with everyday support. More than anything, we are grateful for how openly you shared that perspective with us and how naturally you let us into a part of XetNET’s reality that we would otherwise never get to see.

It is a genuine pleasure to share Erno’s perspective with our community, especially because this is not a story built around pretending AI is magic. It is a story about a good support team finding a better tool, testing it against standards that were already high, and gradually trusting it because the work justified that trust. Seeing Konrado.AI become useful in exactly that way gives us a very real sense of satisfaction, and the prospect of continuing this collaboration with Erno and XetNET makes us even more curious about what the next chapter will look like.

Ready to Transform Your Support?

Save time, reduce costs, and free up your team with Konrado.AI