Key takeaways from Cytora’s AI Forum 2018 Conference

Last month, Cytora hosted it’s biggest event to date — the inaugural AI_Forum. Produced in collaboration with knowledge partners, Boston Consulting Group, the ambition was to strip away the hype and discuss the real applications of AI in commercial insurance, with a focus on concrete economic benefits and practical advice for CEOs.

AI Forum attracted industry leaders across insurance, finance, and technology. Speakers included David McMillan – COO, QBE Insurance, Sam White – CEO, Pukka Insure, Siraj Khaliq – Partner at Atomico, and Cytora’s CEO Richard Hartley.

Richard Hartley opened the event with Cytora’s definition of AI:

“We understand AI as the ability for a machine to make an equivalent level of decision to a human, such as a risk selection decision, based on its ability to identify patterns in data.”


Richard Hartley: AI makes an equivalent level of decision to a human


The time to act on AI is now

During the first panel session, ‘Using AI to enhance pricing’, speakers were in agreement that AI is going to be transformative for commercial underwriting, particularly in SME insurance. Panellists suggested that the SME market is on the verge of becoming fully automated due to the homogenous nature of risks and ubiquitous data.


Siraj Khaliq: The time is absolutely now


“The time is absolutely now,” said Siraj Khaliq, Partner at Atomico (previously CTO at Climate Corporation). “AI essentially enables us to make better predictions based on the data itself telling you something, rather than having a human interpret the data to come up with imperfect rules and imperfect models … the time is absolutely now, and if it’s not now for you, you may not be around to react once you realise that it is.”


Use AI to optimise efficiency and deliver better customer experiences in insurance


During the second session, ‘How will AI change the structure of the insurance value chain?’, panellists were unanimous in agreeing that the way for CEOs to make progress with AI was to identify well-defined, tractable problems within their organisations, and to attack them directly. Improved customer experiences, better pricing, and improving operational efficiency were identified as good places to start.


“There is going to be a revolution. I believe that in the next 12 to 18 months the whole game is going to change,” said Sam White, CEO at Pukka Insure. “It’s consumers that will drive this, every day they are being offered experiences that are far more appealing than what they are currently using.”


Sam White: There is going to be a revolution


These problems should be tackled by combinations of technical and business teams, not hidden away in R&D rooms. By operating in plain sight, technologists learn from the business expertise, and the resulting solutions are more likely to be valuable and adopted.


Insurtech should be seen as an opportunity, not a threat


During an insightful fireside chat, David McMillan, COO at QBE Insurance, discussed his experiences of integrating new technologies at QBE and in his previous role at Aviva.


David McMillan: Insurtech is an opportunity, not a threat


Put simply, insurtechs are great at focused technology development applied to specific problems along the insurance value chain. They aren’t as good at replicating insurers’ established advantages of underwriting expertise, scale, capital resource and customer base. Partnership with insurtechs is a great model for benefiting from the outputs of focused R&D activities without assuming the risk of building the solutions in-house.


“Cytora is really helping us to transform the way we price and the way we underwrite, they are also helping us to understand what the future could look like,” said David McMillan.


For insurance CEOs who accept that disruption is inevitable, the challenge is to ‘win the transition’


Those incumbents that thrive will be the ones that foster a culture of innovation, with a DNA of data, embracing the best that insurtech has to offer.

For more information about Cytora AI_Forum please visit