Lizzy – a chatbot that helps its users to insure other passengers and also answers the most important questions about car insurance.
How did the idea of Lizzy come about?
For companies that have little or no experience with chatbot technology, it is advisable to start with a limited and manageable use case. That was also HDI Versicherung’s idea. There is no one right solution for the first use case, but in general the following criteria should be considered:
- Questions and answers must be predictable
- It must be possible to narrow down the use case
- Questions must appear repeatedly and have the potential to be answered automatically
- The introduction of the chatbot creates added value for the company
HDI Versicherung initially used social listening tools to find out what its customers and target group were talking about. The customer service employees and other colleagues were also asked which customer inquiries or customer processes are recurring and where there is potential for automation. The results of the surveys and observations led to the idea of developing a chatbot to expand the driver base. This use case can be limited very well, the process is always the same and there is a high potential for automation.
How do employees become chatbot fans?
A chatbot project needs internal fans and followers. It is therefore all the more important that the project is communicated openly from the outset and, above all, that employees who tend to have a negative attitude towards the new technology are informed about the opportunities offered by chatbots.
At HDI Versicherung, communication around the chatbot project was very open right from the start and relevant departments and teams were informed accordingly and involved in the project.
How does a chatbot get its personality?
Good chatbots should also have a suitable personality. The digital assistants represent the company and their personality must match both the company’s image and the target group. This template offers a little help in finding your personality and designing your dialog.
HDI Versicherung also took the process of finding the right personality very seriously. Lizzy was created with the help of internal workshops and external research. Lizzy is deliberately female because external research has shown that customers usually prefer women as advisors.
What does Lizzy do well?
Lizzy is a chatbot that works both rule-based and intent-based. When it comes to taking out insurance, Lizzy works on the basis of rules and guides the user through a predefined dialog.
However, when it comes to general questions about car insurance, Lizzy works with the help of artificial intelligence (AI) or intent-based.
Right from the start, Lizzy introduces herself to the user as a chatbot or virtual assistant and also writes that she is still learning. So it raises the right expectations right from the start. The most important thing here is that it pretends to be a chatbot and not a human.
Lizzy realizes when she hasn’t understood something and is very quick to admit it. If it has not understood something, it offers to send the user an e-mail. If the user selects the e-mail option, it records the user’s contact details and request and forwards them internally. This means that no customer inquiries get “lost” and the HDI team can then contact these users personally.
If users want to start small talk with Lizzy, she can react, but tries to quickly steer the conversation back to the main topic. Lizzy writes in a focused way and tries to motivate users to take out insurance.
You can hear even more valuable insights into the development of the Lizzy chatbot and how it can continue in this podcast episode.
And if you would like to chat with Lizzy yourself, you can do so here.
