Fewer tickets: optimising IT service desks with AI

 

The IT service desk has always been subject to cost calculations. With AI, this has become even more pronounced, as many IT and finance managers hope to take the performance and cost-effectiveness of their contact centres to a new level. This is entirely feasible, as model calculations and real-world project experience show: compared to a traditional service desk, personnel costs in an optimised environment are around 30 to 50 per cent lower. In addition, there are indirect financial benefits (if implementation is successful), such as higher productivity and customer satisfaction. However, AI cannot meet exaggerated expectations of fully automated operation.

In our survey on the 2025 IT agenda, we asked IT executives to name the most effective cost-saving measures for times of crisis. In terms of processes, the consensus was in favour of so-called 'shift-left' solutions, which either automate simple service requests or enable those affected to resolve issues directly amongst themselves. This approach forms the basis of the most important optimisation strategy for IT service desks: ticket avoidance. Automating activities is a key part of this strategy, and this is where AI comes into play.

Avoiding tickets in the IT service desk

AI-supported self-service and automation tools include chatbots, online communities and smart knowledge databases. They provide users with immediate assistance, eliminating the need for human intervention. These systems are becoming increasingly context-sensitive and can even act proactively, for instance by sending in-app notifications or solving problems automatically based on usage behaviour. Additionally, predictive AI analyses identify patterns in user queries, providing solutions at an early stage. A tutorial or help page is provided before a ticket is created.

 

Is an optimised helpdesk worthwhile?

However, there are other ways to reduce the number of tickets in the service desk besides AI, such as self-service portals. According to the results of our benchmark projects, if all measures are implemented, the number of tickets can be reduced to around three to five per user per year. By contrast, the average ticket volume for traditional IT service desks is eight to twelve tickets per user per year. In practice, this equates to a reduction in ticket volume of around 60 per cent. Particularly advanced organisations with a high degree of automation and excellent user guidance achieve an even greater reduction, with less than two tickets per user per year.

There are fewer tickets, but they are more complex

A similar trend can be seen in the first contact resolution rate, which can be increased by around 10 to 15 percentage points through specific optimisation measures. Providing targeted support for agents through AI-assisted diagnostic tools also contributes to this improvement. However, the downside is that, since simple queries are resolved in advance through the use of AI and other measures, the IT service desk receives a higher volume of complex tickets. Consequently, individual tickets requiring human support involve greater effort and higher costs; in our projects, we have determined that ticket costs can be 30 to 50 percent higher. Nevertheless, the reduction in the number of tickets often leads to lower overall costs, as the lower volume offsets the increase in costs.

AI in the service desk is part of optimisation

AI could solve the problems faced by service desks, but not on its own. It has great potential for back-end processes, such as pattern recognition and customer chatbots. This fact, combined with aggressive marketing by AI providers, has generated financial interest among senior management. However, AI investments are high, remaining tickets are becoming more complex and, ultimately, automated shift-left solutions cannot be implemented in all organisations and departments, such as the night shift in a hospital ward. Nevertheless, it is still possible to systematically reduce the probability of an incident here.

It is important for companies to develop a sensible AI strategy for IT service centres and arrange all service provision modules optimally. However, those who rely too heavily on AI may negatively impact customer satisfaction. Following the backtracking of Klarna and Duolingo in mid-2025, having previously aggressively marketed their use of AI, it has become clear that the hope for purely AI-driven customer service quickly reaches its limits. The conclusion is that: AI is not a solo player, but a very effective team player when used appropriately.