Adnovum Blog

Conversational AI for insurances

Written by Mark Bosshard | May 10, 2022 3:19:29 PM

The IFZ Conversational Insurance study reveals some insightful results: In Switzerland, almost 50% of the surveyed policyholders have already chatted with their insurance company at least once, around 25% of them even several times. Among those under 18, three out of four have chatted with their insurance company at least once. These changes, which reflect recently accelerated developments in our communication behavior in everyday life, have consequences for insurance companies: the expectation of insured persons to be able to solve their concerns on various channels without friction is constantly increasing.

At the same time, the changes also bring opportunities: Traditional, time-consuming channels, such as e-mails with long introductions and complex cover letters, or call center calls to be answered in real time, can be replaced by live chats, chatbots, and voicebots. If handled correctly, this improves the 360-degree customer view for the insurance company thanks to improved data collection and significantly reduced manual efforts as a result of the high level of automation.

Adnovum pursues two strategies, especially for calls to the call center, which are time-consuming for the company and often long and tedious for customers:

  • Call Avoidance
    Improved app and chatbot offerings are being used to prevent customers from picking up the phone and making a hotline call in the first place.
  • Call Deflection
    If customers do call, an automated response is first played (also available in Swiss German) on the basis of analyzed customer queries. If this is successful, or if the request can be handled online, the calling customer (often from a mobile device) is sent an SMS with a link that redirects the call to an online process (e.g. form) and is thus «deflected». In other words, the call does not reach an employee, but the customer is instead automatically guided (back) to online processes.

To achieve both these call center relief measures while meeting customer needs and 24/7-style reachability on the most popular channels (e.g. WhatsApp), Adnovum combines the voicebot approach with state-of-the-art chat platforms to create a comprehensive solution. Why this powerful combination leads to real added value is easy to see in the concrete application:

Voicebot


A voicebot can be used in many ways, for example:

 ... as an automatic FAQ answerer, which takes the customer's question in 1–3 sentences and then answers it verbally as well as with a further SMS (incl. link to the online process)

 ... as a self-service processor, taking details of damage, for example, and recording simple claims directly and filing them in the claims system.

... as a callback planner, which suggests several time slots for a callback to the customer at peak times of the hotline, and lets the customer select one.

... as an analysis tool for calls with employees, e.g. to determine the distribution of speaking time between customers and employees, but also to carry out more complex tasks, such as the classification of the discussed topics (e.g. sales potential) across large numbers of calls.

Chat Platform

A chat platform solves several problems, for example …

... it enables live chat with the customer across different channels, e.g. from website chat to WhatsApp.

... with a chatbot, which plays out frequent responses in an automated way without the chat getting to an employee.

... by assisting via suggestions to the live chat agent, which, when a choice is made, sends a text template directly to the customer, while further feeding the AI model with each choice made by the employee and thus making the bot steadily smarter.

Does your insurance company also have the potential to use this or similar automation in the area of Conversational AI to reduce employee efforts and at the same time make the customer experience more modern and appealing? We would be pleased to show you the opportunities and risks of possible Conversational AI project setups in a non-binding exchange.

 

[snippet_article_cta id="webinar-AI-en"]

 

[snippet_article_cta id="article-factsheet-en"]