Kömpf24 launches digital employee “KIM” with ElevenLabs

1. Executive Summary

Kömpf Onlineshops GmbH, a leading online retailer for products related to garden, home, and living, faced a critical operational challenge: a monthly call volume of over 10,000 inquiries was being handled by a service team of just three employees. The result was an average wait time of 1.5 minutes, which was frustrating for customers. Through the strategic implementation of ElevenLabs' Conversational AI, a highly intelligent, AI-powered first-level support agent named "KIM" was created. The results were transformative: the average wait time was reduced to just 15 seconds, the overall efficiency of the service team was massively increased, and a digital employee was integrated whose operating costs equate to an "hourly wage" of only €5.48. This case study highlights the journey from the initial challenge to the implementation and the impressive results.

2. The Client and Their Challenge: At the Limits of Scalability

A Traditional Company in the Digital Age

Kömpf Onlineshops GmbH is much more than just a single online store. As a traditional family-owned business with roots dating back to 1934, it now operates a broad portfolio of over 20 specialized e-commerce platforms, including well-known brands like grills.de, mein-gartenshop24.de, and oase-teichbau.de. This multi-shop approach allows for deep specialization but also leads to enormous complexity and a high volume of customer inquiries.

The bottleneck in customer service was clearly defined:

• Enormous Call Volume: With around 10,000 calls per month, telephone customer service was the most frequented point of contact.
• Limited Human Resources: A dedicated but very small team of three people was responsible for handling this flood of inquiries.
• The Critical Result: An average wait time of 1.5 minutes (90 seconds) per call.

This wait time was more than just an unfavorable metric; it was a tangible business risk. Long queues lead to dissatisfied customers, increase the likelihood of abandoned purchases, and place a heavy burden on the service team, which had to primarily deal with repetitive initial inquiries such as order status or simple product questions. A traditional increase in personnel would have been costly and not flexible enough to react to seasonal fluctuations.

3. The Solution: Custom-Tailored Intelligence Instead of a Standard Bot

Kömpf needed a solution that went far beyond a simple, script-based chatbot or a rigid IVR (Interactive Voice Response) menu. The requirement was for an intelligent system that could understand the complexity of the product portfolio, integrate seamlessly into the existing IT landscape, and conduct natural, human-like dialogues. The choice fell on ElevenLabs' Conversational AI.

More Than Just a Bot: A Future-Proof and Adaptable Platform

The decisive advantage of this platform was summed up by Christian Langer, AI-Automation-Manager at Kömpf24: "What we [...] liked was the new Conversational AI from ElevenLabs. [...] that you can now integrate custom LLMs, that you can integrate MCP servers. And that makes the whole thing very powerful."

This statement underscores the strategic core of the decision. Kömpf was not forced to use a generic third-party AI model. The flexibility of the ElevenLabs platform offered crucial advantages:

• Data Sovereignty and Control: By connecting their own LLMs, Kömpf retains full control over its data and the "intelligence" of the system.
• Maximum Adaptability: The AI agent can be continuously trained and refined with its own current company data without having to wait for the development cycles of an external provider.
• Future-Proofing: The system is open to future technological developments and can be adapted to new business requirements or expanded processes at any time.

Thus, Kömpf created not just a solution for a current problem, but a sustainable, scalable platform for the future of customer communication.

4. The Implementation: The Creation of the Digital Employee "KIM"

The successful launch of "KIM" was based on two fundamental pillars: a deep knowledge core and a perfect technical integration that turns "KIM" into a true process manager.

Pillar 1: The Knowledge Core – A 28,879-Word Prompt The success of "KIM" began not with code, but with knowledge. An exceptionally detailed prompt of almost 29,000 words was created. This prompt is the digital DNA of Kömpf's entire customer service knowledge and includes, among other things:

• Detailed product information, specifications, and compatibilities.
• Answers to hundreds of the most frequently asked questions (FAQs).
• Step-by-step process instructions for returns, delivery status inquiries, and complaints.
• Guidelines on tonality and the specific communication style for the various brands under the Kömpf umbrella.

Pillar 2: The Technical Workflow – "KIM" as an Intelligent Process Manager The call process was designed to be seamless for the customer while complex processes are managed in the background:

1. Call & Routing: A customer calls and is routed to the AI via Twilio.
2. AI Activation & Dialogue: "KIM" starts the conversation and analyzes the customer's request in real-time.
3. Deep Process Integration and Categorization: This is where "KIM's" true strength lies. Instead of just recognizing keywords, the AI understands the context and initiates specific actions. Based on the customer's request, it precisely distinguishes between:
4. • Consultation: Determines whether a general question can be answered by the AI or if a transfer to a human agent is necessary for complex product advice.
   • Delivery Date: Can independently track shipments or forwards specific inquiries directly to the logistics department.
   • Complaint: Differentiates between a new complaint, for which a ticket is created, and an existing complaint, where the status is checked and the case is assigned to the correct ticket.
   • Supplier Inquiries: Recognizes calls from freight forwarders and can route them to the appropriate contacts in the warehouse.
5. Automated Backend Processes: Based on the categorization, "KIM" acts autonomously:
6. • Ticket Creation in Zendesk: Automatically creates a new, correctly categorized ticket.
   • Data Validation: Checks order numbers against the backend in real-time.
   • Intelligent Escalation: Forwards only those cases to human experts that require their specific knowledge – including all information already collected.

5. The Results: A Transformation in Measurable Numbers

The results of the implementation were immediately noticeable and exceeded expectations.

Metric 1: Drastic Reduction in Customer Wait Time by 83%

• Baseline (March 2025): 1.5 minutes (90 seconds)
• Final (June 2025 - AI on 3 Hotlines): ~0.25 minutes (15 seconds)

Metric 2: The Outstanding Cost-Effectiveness of the AI Employee The economic viability of the solution is equally impressive. The operating costs for the AI agent "KIM" amount to approx. €5.48 per hour, based on an operational time of 12 hours a day, 365 days a year.

Metric 3: Empowerment of the Human Service Team An important qualitative success was the relief of the human service agents. Freed from the constant flood of repetitive initial inquiries, they can now apply their expertise to solving complex, consultation-intensive, and value-adding customer issues.

6. Conclusion: A Future Model for E-Commerce Customer Service

The collaboration between Kömpf Onlineshops and ElevenLabs is more than just a technological success story. It is a strategic milestone that shows what the future of customer service in e-commerce looks like. Through the intelligent combination of a flexible and adaptable AI platform, a deep, company-specific knowledge core, and seamless integration into business processes, Kömpf has achieved three core objectives simultaneously:

1. Maximizing the Customer Experience: Frustration from long wait times has been virtually eliminated.
2. Increasing Operational Efficiency: The human team is optimally used as an expert authority.
3. Sustainable Cost Optimization: A scalable and extremely cost-effective service channel has been established that can grow with the company.