So far, there is no single solution for integrating chatbots quickly and effectively into a company in the long term. Observations show that every company goes its own way and that is a good thing.

Different approaches can be chosen depending on the degree of digitization, resources, budget and target groups.

What they have in common, however, is that every chatbot project should be viewed as a learning project. This means that the results must be closely monitored and used for further development and optimization.

AXA opted for chat as a communication channel very early on

AXA Switzerland launched its first live chat in 2016. Live chat was offered on selected websites or in the customer portal and was primarily intended to help customers who were unsure about the digital self-service offers. This was the first time that the company, its employees and customers were able to “get to grips” with chat communication.

Since 2017, it has also been possible to send customer feedback to AXA Switzerland via Twitter private message or Facebook Messenger and even receive an individual response from an AXA colleague.

In 2018, AXA was one of two Swiss companies to be the first to introduce Apple Business Chat for its customers. iphone customers now receive help directly via iMessage.

AXA Switzerland’s first chatbot also saw the light of day in the digital world of web browsers and Facebook Messenger in 2018 . This first chatbot was able to guide customers affected by glass damage to their car through the entire claims process without provoking a channel change.

AXA then introduced the principle of expert and master bots in 2019. A master bot was developed with the virtual assistant Ada, who receives all chatbot requests on the website. Using artificial intelligence (AI), Ada understands the user’s intention in most cases and can then forward the user to an appropriate expert. If the expert already exists in the form of an expert bot, Ada technically forwards the conversation to the expert chatbot in the background, but the interface remains identical for the user and they receive their information directly from Ada. If there is not yet an expert in chatbot form, the conversation is forwarded to an employee with the relevant knowledge.

Thanks to the principle of expertbots, AXA is able to expand the chatbot’s expertise in a short space of time. This is because each specialist department can create its own expert bots and link them to the Ada master bot.

Chatbots complement self-service forms

The digital experience team at AXA Switzerland has learned that chats and chatbots in particular are a good addition to the classic self-service forms. In many cases, insurance still requires advice or at least a dialog. Some customers do not feel comfortable if they can only calculate their claim or premium in a self-service form. At the same time, the desire for 24/7 availability is increasing and hardly anyone is happy to wait in a telephone waiting loop.

Chatbots offer an ideal bridge here. They are available around the clock and offer customers a solution to their problems through dialog.

Chatbot as a central self-service touchpoint

In the long term, AXA’s digital assistant “Ada” is set to become the central self-service touchpoint. At the same time, customers should not be denied personal contact via the other chat, telephone or offline channels.

Customers or potential new customers who chat with Ada always have the option of asking for a human contact. If employees are available during normal opening hours, Ada offers the live chat channels mentioned above. If this is not possible, it takes the contact details and the user is contacted personally as quickly as possible.

Existing customers are also informed that they can write to their advisor quickly and directly via text message using the Myaxa app.

And how is it used?

The figures show that the AXA team is not wrong in its efforts towards live chat and chatbots. In 2020, there were already almost 43,000 chat messages and the numbers are clearly on the rise. The usage figures for the chatbot Ada are also particularly pleasing. These show that the dialog-driven self-service channel in particular is gaining more and more acceptance (a good 11,000 chats in 2020).

You can read even more background information on AXA’s chatbot strategy in this article and even listen to it in the corresponding podcast episode.