Select your region:
Select your language:
English

Get Ready for Chatbots! - Part 5

Doing Chatbots Right

In our previous articles we have learned that AI-based chatbots need training to become better at natural language processing. Hence, a well thought-through strategy is necessary to efficiently collect and label the required training data. Additionally, since chatbots are used for direct customer interaction, they need to be introduced properly. In the following, we will describe an approach that makes this possible.

How to Not Introduce a Chatbot

Insurers are still hesitant about investing too heavily into chatbot technology. Before committing more resources, they first want to assess the reaction of their customers. They typically start experimenting with chatbots by initially automating only a few selected processes with customer interaction. Accordingly, the customers of the insurer need to be informed beforehand that they will only be able to access a very limited number of services when using the chatbot. Any customer request that lies outside the scope of the chatbot is answered by a response such as «Sorry, I could not understand you».

However, there are two problems with this approach:

First, chatbots with limited scope will only gain little acceptance among customers due to the poor user experience: customers must constantly make the mental effort to consider how to phrase their messages such that the chatbot understands their request.

The second problem is that these chatbots are deployed to work without human involvement on the insurer’s side. Hence, they need to be 100% accurate in processing user input to avoid errors in the business process. However, attaining 100% accuracy often requires a lot of training for the chatbot, and in some cases it may even be impossible to achieve with current technology.

Gradual Automation

We therefore suggest a different approach in which human agents and chatbots work side by side to achieve an optimal customer experience (Figure 1):

  1. First, introduce chat (without chatbot) as a new interaction channel through which customers can access the full array of available services. At this stage, incoming requests are still processed by human agents, typically in traditional customer service centers.
  2. Next, a basic chatbot is deployed into the chat channel established in the previous step. This chatbot will only be able to handle simple requests initially. Whenever it receives a request it cannot process, it automatically hands over control of the chat to a human agent who will then reply to the customer.
  3. After its deployment, the training data for the chatbot is gathered on-the-fly by continuously collecting the customer requests and the corresponding decisions of the human agents. With time and training, the chatbot can handle more and more request without human intervention.
  4. The data gathered for the chatbot can be used to further expand its capabilities. For example, it could be utilized to extend the chatbot to speech-based interaction, or to generate responses based on customer analytics and sentiment analysis.



Figure 1: Gradual automation of customer interactions via chatbots

Chatbot-Human Handover

An important requirement of the gradual automation approach is the smooth transition between chatbots and human agents. Unless general artificial intelligence steps out of the realm of science fiction, there will always be times where a chatbot needs to escalate to a human being.

Whenever the chatbot is not sufficiently confident about the processing of user input, it sends its proposed response to a human agent (Figure 2). The agent then decides whether to directly forward the chatbot’s response to the customer, or to modify the response first before sending the response. The agent’s decision is then used to train the chatbot for similar future requests.


Figure 2: Chatbot - Human Handover

As you can see, the introduction of a chatbot into customer interaction requires a cleverly devised strategy to be successful. By using the suggested human/chatbot hybrid approach, insurers can train and make use of chatbots immediately while offering their customers access to all services at all times. In our next and last article of the series, you will catch a glimpse of the future of chatbots: we will talk about virtual assistants, voice-based interaction and the use of customer analytics.


You might also be interested in the following articles:


Are you curious about trying out our prototypes? Would you like to talk more about chatbots? You can reach us directly, using the contact details below.

  • Dr. Dominik Langer
  • Weili Gao
  • Karim Attia
Dr. Dominik Langer
Associate Partner
dominik.langer@synpulse.com
Dr. Dominik Langer
Weili Gao
Consultant
weili.gao@synpulse.com
 Weili Gao
Karim Attia
Consultant
karim.attia@synpulse.com
 Karim Attia
Your browser is out of date!

Update your browser to view this website correctly. Update my browser now

×