Creating a new, AI-powered sales channel

The Challenge

Helvetia, a leading Swiss insurer, faced increasingly high customer acquisition costs. Adding sales headcount risked cutting into already slim margins, with the cost of smaller policies sometimes exceeding insurance premiums.

However, cost savings shouldn’t mean cutting corners or sacrificing customer experience. With this in mind, Helvetia sought to automate conversations and engage users proactively, without sacrificing quality. The Swiss insurance industry sets the benchmark for data privacy, another absolute requirement.

The Solution

Helvetia decided to build a powerful AI assistant that could offer insurance policies to customers via text message, with Rasa providing the underlying NLU and dialogue management. The team was able to design an experience that considers the user’s context and allows the customer to engage using natural language.

Rasa Enterprise’s graphical user interface allowed non-technical team members to improve the assistant and train it with real user data.

Additionally, Helvetia’s lead developer on the project was able to tweak the machine learning models specifically for their use case, a customization that resulted in improved accuracy. The assistant was deployed safely on-prem, fulfilling data privacy requirements.

The Results

The careful implementation and ‘go-live’ interactions with real customers produced incredible results, achieving a conversion rate of over 30% via text message.

Using Rasa Enterprise, the team was able to leverage a wider range of Helvetia employees, including those in non-technical roles, to establish excellent customer experience and iterate faster on feedback. The company now plans to expand the AI assistant to more use cases.

Rasa Enterprise allowed us to significantly reduce the engineering costs of our conversational AI projects. Being in control of our data and where Rasa is deployed is very important for us.

Florian Nägele, Project Manager at Helvetia