RTL+ in an Interview: Using AI to Create a Streaming Buddy and a Personalized Customer Experience

RTL+ has grown rapidly in recent years, from around 250,000 to more than seven million paying customers. This also raises the demands placed on customer service, as streaming content such as sports events or reality formats often airs outside traditional service hours. If an issue occurs during a live event, customers need immediate support so they don’t miss any important content. Traditional chatbots quickly reach their limits here because they are often too one-dimensional and can only answer simple standard questions.

RTL+ customer care therefore deliberately opted for a Salesforce Agentforce pilot project. The AI agent is integrated into WhatsApp and provides customers with fast, personal support around the clock. WhatsApp was chosen because many viewers already use the channel as a second screen while streaming on their TV.

Christian Achatzi, Partner for Media at Salesfive, spoke with Christiane Konrad, Vice President CRM & Customer Care at RTL+, about the project.

Christian Achatzi: Christiane, why was Agentforce the right choice for you compared to traditional bots?

Christiane Konrad: We have already had very positive experiences with AI-based chatbots, especially when it comes to automating standard requests. The next step was clear for us: to further develop our existing automation toward a more personalized customer experience that also covers more complex issues. With Agentforce, we are deliberately enhancing our existing solutions with precisely these capabilities and integrating them more deeply into our system landscape. For us, this is the consistent next step in the evolution of our AI-powered customer service.

Christian Achatzi: Agentforce operates much more autonomously than earlier systems. How did you experience this new type of software during the project?

Christiane Konrad: An AI like Agentforce is in fact more like a new colleague with clearly defined guardrails and continuous training than a traditional piece of software. You no longer rigidly program everything in advance; instead, you define guardrails. In our pilot, we initially experimented: sometimes we gave the agent too much freedom, then we tightened the constraints again to build confidence. It’s a continuous test-and-learn process. It was important to us to take away employees’ uncertainty. They need to understand that they can steer the AI.

Christian Achatzi: How do you ensure that the quality of the AI’s responses meets your high standards?

Christiane Konrad: Just like we do with our human colleagues in the contact center, we run continuous quality monitoring. We use AI to analyze the AI and make patterns visible, because manually reviewing that volume of chats would hardly be feasible. The reporting capabilities in Salesforce give us a solid initial orientation and help us identify relevant cases in a targeted way. At the same time, it’s important to us to critically challenge these results and benchmark them against our own KPIs. This is how we continuously evolve our monitoring.

Christian Achatzi: A key issue with AI is always cost. Is the business case already paying off for you?

Christiane Konrad: Absolutely – we’re monitoring the business case very closely. One major lever was reducing the handover rate. In the beginning, the agent often handed over to a human when it wasn’t confident. By optimizing the prompt, we’ve already been able to cut these handovers by more than half. That has a direct impact on costs and first-contact resolution.

Christian Achatzi: You’re pursuing the vision of a streaming buddy. What does that vision look like?

Christiane Konrad: We want to create a digital companion that doesn’t just react, but in the long term provides proactive support, offers recommendations, and solves problems. It should be able to detect signals independently. For example, if streaming stutters or a customer can’t find something, it immediately offers help. The buddy should know our customers, understand what they need, and thus prevent churn and increase lifetime value.

Christian Achatzi: To wrap up: How did Salesfive support you on this journey?

Christiane Konrad: Working with Salesfive was extremely close and truly collaborative. What I found particularly valuable was that Salesfive also actively challenged our ideas. In a dynamic environment like AI projects, it’s precisely this honest, expert exchange that is critical to achieving lasting impact. Overall, we rely on a strong partner ecosystem to optimally combine different strengths.

• Efficient solution: Reduced the agent-to-human handover rate by over 50% through prompt optimization.
• 24/7 Support: Available via WhatsApp, especially during peak periods such as sporting events or at weekends.
• Complex requests: The WhatsApp channel is increasingly being used for more in-depth queries and is gradually replacing traditional channels such as email.
• Scalable quality: Use of AI-powered monitoring to ensure a high-quality customer experience.