New research conducted by bot training company INTNT.AI reveals that less than 7% of the UK’s top insurers have a chatbot.  Is the insurance industry missing or ignoring the chatbot revolution?

AI is set to take over the world. Although that may have been the premise of a dystopian sci-fi film of the past, it is true of the present, although with much less sinister undertones. Rather than sentient, malevolent bots seeking to destroy humanity, AI instead holds the answer to increasing demands on customer service.  Bots powered by artificial intelligence are today leading a much broader digital transformation of the web across all industries. What is surprising then, is that only 6.75% of the websites of the top 148 companies listed by Insurance Business Mag have a chatbot. This begs the question, why there is such a low adoption of chatbots within the insurance industry? Perhaps more crucially, what are the digital transformations in customer experience that they provide?

If you are completing a purchase on the web and want some more information on the product who do you ask? Similarly, if you are a customer interested in purchasing insurance online and want to raise a query about the policy what do you do? Traditionally you would seek to contact customer service and be presented with either an email address with no known timescale for the reply or a phone number. But who likes to be put in an hour-long queue for a customer service representative, despite your phone call being “important to them?” These are the initial leads that could be quite easily handled by a chatbot.

However, the majority of the insurance industry is yet to enjoy the cost and time-saving benefits of AI technology. This technology and chatbots themselves were supposed to have a glorious future. Indeed, Business Insider predicted back in 2016 that 80% of businesses would adopt a chatbot by 2020, a prediction that has proved to be catastrophically wrong. Now chatbot usage is a grim reality, the wondrous AI-powered future has been plagued by statements of “Sorry I don’t understand.”

Missing or Ignoring?

In secondary research collated by INTNT.AI, the top 148 insurance companies in the UK listed by Insurance Business Mag were surveyed to check for the presence of a chatbot. Of these, only 10 companies were shown to make use of AI to help with their customer service requests. Of the 10 that did have a chatbot, 70% of them were freeform, meaning that the AI can respond intelligently to any sentence that the customer types, with the other 30% being scripted, meaning that the customer is presented with several multiple-choice options to describe their query. Interestingly, of those that didn’t have a chatbot, 7.81% of them had a live-chat-only function, and 24.2% of them had an online contact form.

This surprisingly low adoption rate potentially has a big negative impact on insurers, as customers are generally directed to an agent through live chat, or with an agent having to respond personally to an online contact form. In consequence, costs of processing enquiries goes up, customer experience goes down as, despite the cheery music when you are placed on hold, no one is happy to wait for support that could be instant.

The takeaway? The insurance industry is either missing or ignoring the chatbot revolution that is transforming conversation across the web.

Why do companies need a chatbot?

The motivation for chatbot adoption doesn’t need to be just outweighing the negatives of intensive customer service support, no matter how lucrative this may be. Chatbots also provide a much more tailored and conversational response to potential customers instead of just being greeted by a standard web page. Indeed, in research by Userlike, 68% of respondents felt that they had a positive experience with a site as a result of the chatbot being able to answer their query quickly.

It is a fast, easy and convenient way for users to clarify their questions and complete their purchase. This is particularly useful during non-working hours which is when most browsing occurs. Specifically for the insurance industry, it helps users to self-qualify in terms of age, medical pre-conditions, etc. This saves time both for users and insurance agents that would have to answer these questions if the bot was not available.

This shows that, for customers, chatbots can provide a useful initial point of contact with the company outside of intensive customer service channels that can help start a conversational customer journey and thus leading to more sales. This itself was found during a study by Juniper Research who estimate chatbots will drive $112 billion worth of eCommerce sales by 2023.

What went wrong? 

2019 saw a frenzy of interest and enthusiasm for artificial intelligence.  The year posted the greatest number of AI start-ups to date, and industry commentators heralded the era of the machine had arrived.  During this period of excitement about the potential of technology to save the planet and deliver world peace, Business Insider confidently predicted that 80% of businesses would be operating chatbots by 2022.

From our research this month, that clearly hasn’t happened.  Sadly, the experience of using chatbots 3-4 years since their arrival has been worse than expected. Perhaps this is the reason why adoption is so low.  Ian Tomlin, Managing Director of INTNT.AI, agrees.  “The grim reality is that untrained chatbots suffer the same problems as untrained customer service staff.  They lack the fundamental understanding of what customers want and misread their intentions.  Providers, faced with a long and costly learning curve, stick to what they know, even when it looks nothing like the AI future they were promised.”

The glimmer of hope

The scarcity of chatbots within the insurance sector and business more broadly shows that there is a large amount of skepticism towards the use of AI in customer service channels, and rightly so. Indeed, in the past, the main problem that has plagued chatbot use has been the proportionally large number of misread intents, which typically hovers around 35%. This can create obvious frustration for the customer and start the conversational journey off on a bad footing.   However, preconceptions about untrained chatbots are outdated and AI technology has caught up with its lofty expectations. The dark ages of chatbot usability are finally beginning to dwindle.

The next generation of trained bots can understand your intent

Start-up INTNT-AI, has developed a bot ‘auto training’ platform that enables companies and bot developers to improve the accuracy of their bots to understand the intent of what prospective customers and enquirers want using a blend of linguistics and AI; displacing manual training and its long lead time. In one case study, INTNT-AI improved bot accuracy from 37% to 86%. To achieve these results, INTNT-AI automates the bot training process, feeding in chatlogs monthly, and outputting recommendations that can be adopted with a single mouse click. This analysis and recommendations process includes the auto-detection of false positives, false negatives, and clustering new intents for better recognition. Bots become 102% more accurate in just 3 weeks, and 180% more accurate in 8 weeks.

How can this help insurance companies and customers alike? Apart from the obvious time and financial savings that effective chatbots provide, chatbots allow customers to kickstart a conversation at any time. Customer growth and increasing sales naturally follow. Thanks to advances in the science of linguistics and the application of AI, insurers today can reset their preconceptions, and see beyond the fallibilities of today’s generation of untrained chatbots.

The question is, will they take it?

Ian Tomlin
I'm the CEO of the conversational marketing agency and consultancy Newton Day. We help businesses to tell their story and curate profitable conversations with their audiences acting as their internal marketing team.