Has Chat-GPT taken InsurTech Full Circle?
![Group 538-3](https://685080.fs1.hubspotusercontent-na1.net/hubfs/685080/Group%20538-3.png)
Written at Jul 31, 2023 2:24:08 PM by Dave Connors
At the end of my “corporate” career, and in the early days of distriBind I would go to Insurance / InsurTech conferences and hear talk about “AI” and how it was going to “Transform” Insurance. The vast majority of the presentations left you feeling a bit “Emperor’s New Clothes”: it often boiled down to an Engineer with no Insurance experience had built a chatbot.
I’d often wonder, would any other industry let at outsider stand on stage and tell them how bad they are at pretty much everything, how they’re letting their customers down, and their solution to the problem was giving your customers a stilted awkward text-exchange with a chatbot: often one that had scraped all your social media and knew you’d just booked a skiing holiday and might need extra travel insurance… Weird, and creepy. (Perhaps not unlike the Engineer presenting…)
At a conference in Paris in Autumn 2019, I was out with some of the other start-ups one evening when I put my foot it in by saying exactly this: “AI will have a big impact on insurance, some nerd who’s never worked in insurance and has built a chatbot, won’t” when I realised I was sat next to someone who fit that description to a T. (Needless to say, while I’m still in touch with a few people I met for the first time at that conference, that guy never spoke to me again!)
I could never shake the feeling that this was a very tepid (and unrealistic) view of the applications of AI in Insurance. The use cases around data processing – for ingestion, analytics, and cleansing – were much more exciting to me but were very much in the minority. There were always many chatbots for every Analytics tool.
It’s understandable: right now, the hype and the easiest understood and most public facing applications of UI are in Generative AI – Large Language Models like Chat-GPT, Image generation tools like DALL-E and Midjourney, and the Deep Fake video tools. These are exciting tools for people to use, and easy for anyone to understand the use case.
But this has led to a new proliferation in InsurTech of “AI” meaning “chatbot” (and its fancy cousin the virtual agent). The trifecta is Chat-GPT with text to speech and video generation creating a chat-bot you actually have to speak to. Now, I was born in December 1980, making me a crossover between Gen-X & Millennial (the insulting term for it is Geriatric Millennial). My absolute most Millennial characteristic is an aversion to using the phone – if you’re going to make me speak, it sure as hell better be to a person! And from working with later Millennials and Gen-Z, I can assure you the phone call is not going to be making a sudden comeback. We will have to look elsewhere for the game-changing use of Chat-GPT. Even in text-based chat agents, well Chat-GPT is definitely awesome, but you don’t have to spend long doom-scrolling on LinkedIn or Instagram to see a post trying to teach you how to write “effective prompts” for Chat-GPT. This suggests that despite the wonderful capabilities of Chat-GPT, it’s not yet ready to stop most people simply typing “Agent” as soon as the chat window opens.
As with before the generative AI revolution, the really exciting use of AI remains in data processing and analytics, and for Chat-GPT and the like, it’s in particular extending existing use-cases of natural language processing. For systems applying rules like distriBind, this might be combined with OCR to support automated configuration based on contract wordings/terms; for others it might be extending existing sentiment analysis tools or claims fraud detection. I’m particularly interested in seeing how it might unpick complex contract wordings and provide an easier way to model differing limits, premium allocations, aggregations etc. (You can tell I used to work in reinsurance, huh?) The game changing use for AI in insurance could be any or all of these things. It won’t be a chatbot.
If you're looking to leverage the power of AI in insurance and explore innovative solutions like distriBind, we'd love to hear from you. Contact us today to learn more and discover how distriBind can revolutionise your insurance processes.
0 comments