Customer description‍

A medium-sized IT service provider with its own help desk for customers in healthcare, logistics and government. The support department processes hundreds of tickets every day, mostly around standard questions or recurring issues.

Challenge‍

The support team was under pressure due to a high volume of tickets, many of which were repetitive. This led to longer waiting times, customer frustration and a high team workload.

Solution‍

A solution was developed that is an AI co-pilot based on generative AI (ChatGPT) that automatically generates suggestions for incoming tickets. The assistant has been trained on previous tickets, internal knowledge bases and product documentation, and is used as the first filter in the support process.

Approach

1. Build data analysis and training set
   We collected thousands of historical tickets and labeled them by topic, solution direction, and customer type.
2. Integrate model with ticketing system
   The AI assistant was linked to the existing ticket platform and works in real time when opening new notifications.
3. Validation and fine-tuning
   In the first few months, the output was checked by employees, after which the model was automatically allowed to respond to standard questions.
4. Monitoring and learning
   Employee and customer feedback was used to continuously improve responses and add new topics to the knowledge base.

Results

• 55% of the tickets were handled fully automatically
• Shorter waiting times and higher customer satisfaction
• 30% time savings for support staff
• More focus on complex or critical questions

Learnings‍

The use of a generative AI assistant provided structural relief for the support process. By automating routine tasks, space was created for quality, speed and deepening customer interaction.

‍