AI in the City - event recap and thoughts
Last week we attended AI in the City, an event for AI insurance thought leaders and executives. I wanted to share with you a quick recap of the event and some of the thoughts that I took away at the end of the day.
After overseeing the delivery of numerous AI projects in the insurance industry, each with its own challenges, I have decided alongside my other co-founders who were sharing similar challenges(James Breeze -AI Lead and Steven Walden - Global Head of Underwriting Management at AXA XL, David Carruthers - Home & Pet Change Business Partner at RSA, Craig Civil - Head of Data Innovation at Lloyds of London), to start a community-driven Insurance event with the objective of sharing pragmatic insights in the delivery and adoption of AI.
I had the pleasure of opening up the event, after which James Birch and Alan Tua from Ki Insurance, spoke about Ki - the first fully digital algorithmically driven Lloyds of London syndicate - a disruptive solution they built in only 10 months, which is moving away from the current standard of a document-led solution to a data-led one, based on machine learning models.
Gero Gunkel from ZCAM (Zurich Insurance) spoke about his learnings from deploying several AI projects, and why we need to make AI ‘boring’. According to him, most of the insurance world is stuck in the experimental R&D world. He therefore challenged the attendees to change their mindset from ‘fancy and new’ to ‘fast and predictable’.
Our CEO, Brad Cordova then outlined how we aim to provide the world’s first guaranteed service for insurers. He asked the audience to work together towards reducing the 87% fail rate of AI projects to 0%. The main reason he attributed to the failure rate is the complexity and fragmentation of the ML world and the lack of a good way of combining different sources of intelligence. He then spoke about our mission of making AI easy and how our platform is enabling that.
Last but not least, Pamela Negosanti, VP of Insurance at Expert.ai, spoke about using NLU for processing slips. She outlined the complexities of language understanding (particularly for complex insurance documents) and moving from data to insight. Following a live demo, she outlined some of the benefits of using AI for slips processing, the highest being a 40%-60% reduction in overall time needed.
It was great to see some familiar faces from previous events, along some new faces. While the current pandemic has stopped us from meeting in person, it was great that it allowed us to move the event online, therefore welcoming people from various parts of the world.
I second my co-founder James Breeze's main takeaways:
- Co-creation and collaboration are critical in getting projects live
- We need to make AI repeatable and predictable (boring) and prioritize speed over perfection
- We must be clear of the value we provide and deliver it fast
To that, I can only add that it was great to see some great questions coming from the audience and a great interest in all the topics presented.
Finally, if you’d like to have access to the recording of the day and join the ‘AI in the City’ community, please opt-in here.