The lasting impact of COVID‑19 on commerce and trading is an issue for the future. While old routines may quickly resume as restrictions lift, there may be lasting changes to the ways we get things done.

The pandemic impact showed the importance of remote access to get information and to conduct transactions differently. Customers became self-servers using automated solutions that replicated or replaced some face-to-face interactions.

It’s hard to imagine a lockdown without online and do-it-yourself services enabled by technology solutions. The experience we’ve had as an industry has helped businesses understand what tech solutions work well or less well. Meanwhile, one outcome is that newly enfranchised consumers may prefer to remain empowered. Chatbots have been around for a while, so could this be their moment? And if so, what does this technology offer and can more insurers benefit?

Spixii is a technology company providing this type of customer-facing, automated-insurance solution to Life/Health and P&C insurance companies. To find out more, I spoke with Spixii founder and CEO Renaud Million.

RC: How does a chatbot operate?

RM: A chatbot is essentially a software program with a conversational interface that is usable by either a voice or by typing using a keyboard. A chatbot works best when it has a defined goal and it will guide a user towards reaching this target.

RC: What makes a chatbot “intelligent”?

RM: With a chatbot, and systems with AI solutions more generally, intelligence refers to the ability to achieve these goals. The goals are defined by the chatbot owner, and the machine’s ability to achieve them is based on anything from a simple set of rules to highly complex algorithms.

RC: What type of advanced analytics and conversation insights make a successful chatbot?

RM: Chatbots generate a lot of data, much more than traditional digital tools - such as web forms, which rarely capture the user interaction between screens separated by the “next” and “submit” buttons. The data generated by chatbots are related to the execution of the underlying process and how likely it is going to reach the defined goal. More granular data is generated for the conversation itself, such as where people are stuck, at what point they drop the conversation, on which questions people edit their answers, where more information might be needed and where questions are not clear.

RC: What are the pre-requisites of an enterprise-ready chatbot?

RM: An organisation just embarking on digital transformation starts with the desire to improve a single process, or even to create a new one from scratch. Once settled on a process, the chatbot needs to be integrated into middle and back-office systems. This is a pre-requisite for an enterprise-ready chatbot to deliver value. Capturing, validating and transmitting the data in a secure way to core insurance systems will deliver savings and efficiencies at enterprise level.

RC: Where in their processes can Life and Health insurers deploy chatbot technology effectively, and what is the business case for doing so?

RM: A chatbot can assist with many processes; for example, quote and buy, policy administration, submitting a claim or asking for a pre-authorisation in a health product. The process of buying is complex and, as a result, often carried out over the phone. While these analogue conversations are great, they’re just not scalable. Online digital web forms are too rigid and lack the conversational aspect. Additionally, despite existing tools becoming less effective and the ever-increasing costs of expanding call-centre capacity, the focus is always on the top-line. This may explain why some insurers struggle with digital transformation by not prioritising it on the short-term. Chatbots are helping solve this conundrum from both a technical and regulatory perspective since they can be deployed rapidly and audited.

RC: Do you think people’s response of embracing remoteness during the pandemic furthers the argument for using chatbots?

RM: People were already accustomed to doing things by self-serving online. Pandemic restrictions simply accelerated this - across all industries. Remote working also forced more people to go online for their insurance needs. As more people travel less, the need to physically meet with an agent is disappearing. Consumers are trying first to see if they can do what they need online and then call an agent. On the other hand, while insurers want to do more and be more efficient remote working doesn’t automatically equate with service resilience if a surge in demand occurs. Offering digital-first communication tools - such as a chatbot - can bring both efficiency and resiliency.

RC: What is the next technical development for your chatbot technology?

RM: We have already helped several companies integrate chatbot technologies to their quote and buy, policy administration and claims submission processes. We are now refining these for specific lines of business to help them to grow their portfolios. We also realise that business reporting is critically needed for various departments - ranging from IT, marketing and distribution to operations and claims - but the precise data needs vary for each one. So we built an automated reporting function to bring the relevant data to the correct operational area at the right time. We believe this helps individual units make better informed decisions for the whole business. From a tech perspective, we aim to extend the front-end integration of our chatbot, including more channels and platforms, making the configuration of the chatbot even more accessible and friendly. Also, Spixii has achieved accredited ISO 27001 certification for information security, and our work has been recognised by analysts from Gartner, Forrester and Business Insider.

RC: Will you share some working examples?

RM: We have worked with Zurich Insurance group in the UK since 2018, and together, we automated the first notice of loss for Home and Motor insurance, helping them to win the British Claims Awards in 2018 for best use of technology. We also worked on the quote and buy processes with both the UK Post Office on travel insurance and Bupa’s Private Medical Insurance. To be honest, it is an honour to work on impactful processes that bring a better customer experience and generate a positive impact on operations.

RC: Finally, what is the next step for chatbot technology in your opinion?

RM: Although adoption started some years back, many implementations are yet to come. A lot of companies recognised the operational limitations of call centres but only a few companies recognised the limitation of web forms. The ones that did recognise these two points also understood that these two channels should be supported by a third one that keeps the conversational aspect and yet manages to codify it in a digital way. We see a strong appetite for serving the customer better with the digital experience they expect. Digital functions are gaining more authority within insurance companies. Both forces should lead to more and more successful use cases for chatbots to support the digital transformation of insurance companies. From the technological perspective, chatbots are needed for growing more sophisticated in the analysis of data collected - with a particular focus on psychometric and customer behaviour analysis - but also they need to build a stronger understanding of obligations and duties to keep data protected and anonymized, which is an intriguing challenge for this new wave of data collection.

Renaud Million has worked in the insurance industry since 2010 and co-founded Spixii in 2016. Prior to Spixii, Renaud worked as an actuarial consultant for EY in London. He holds a Masters in Operational Research & Management Sciences from Lancaster University, a Masters in engineering from Arts et Métiers ParisTech, and was named young actuarial professional of the year 2017 by Insurance ERM.