How Did the Schmidt Group Reduce Its Support Incidents by 22% Using AI?
active users per month
monthly conversations
incidents over two years
The challenge: reduce the workload on support staff and give field teams more autonomy
70% of support requests provide no added value
As its international network continued to grow, the Schmidt Group identified a structural imbalance in the management of its support operations:
- Channel overload: 70% of inquiries were general support requests, not technical issues. This workload was exhausting the teams and preventing them from focusing on complex cases.
- Loss of on-site productivity: In-store salespeople were wasting valuable work time searching for answers, at the expense of their core business activities.
- International requirement: The project had to meet the needs of a multilingual network—in French, English, Spanish, and German—without compromising service quality.
The goal: to restore autonomy to field teams, refocus support expertise on high-value-added cases, and control operating costs on an international scale.
The solution: two AI agents deployed for two distinct populations
Woody for the sales network, Nano for headquarters
To precisely meet the needs of each population, the Schmidt Group has structured its project around two complementary agents, deployed on the Wikit Semantics platform:
- Woody (sales network): Integrated directly into the sales team’s sales management tool, Woody answers their business-related questions in real time and automatically creates tickets via the IWS connector. It operates natively in four languages to support the entire international network.
- Nano Agent (Headquarters Staff): Accessible via Microsoft Teams, Nano provides internal support for headquarters teams with the same level of integration and ease of use.
- A gradual evolution: Launched in 2022 within the IT department, the project reached a key milestone in 2025 with the integration of generative AI, before expanding to customer service and international deployment.
The results: fewer incidents, greater autonomy, and increasing adoption
A Steady Decline in Incidents Despite Network Growth
The effects of the deployment have been documented and are long-lasting:
- -10.7% decrease in incidents between 2023 and 2024, followed by an additional -12.4% decrease between 2024 and 2025, despite the continued opening of new stores.
- 2,000 active users per month and more than 3,000 monthly conversations handled autonomously, with an upward trend.
- A structural time savings for support teams, allowing them to refocus on complex, high-value-added cases.
- An enhanced user experience for salespeople, who can instantly access the right information without interrupting their sales activities.
The Schmidt Group's Perspective
Testimonial from Alexandra Colin, Project Manager, Schmidt Group:
Wikit was able to meet our two distinct needs by providing us with two customized chatbots that integrate with our IWS ticketing system. We’re thrilled to have chosen Wikit for our chatbot projects!
Key Factors for the Project's Success
The Schmidt Group Shares Its 5 Recommendations
Drawing on this experience, the Schmidt Group shares five recommendations for a successful large-scale AI deployment:
- Define a clear functional requirement before choosing a technical solution.
- Anticipate organizational changes to facilitate adoption.
- Making the Most of the Testing Phase: The Quality of Testing Determines the Quality of Deployment.
- Entrust governance to the business teams, not just the IT department.
- Make data security a key consideration from the very beginning.
