InsurTalk: Robots are coming – how to save the underwriter from extinction? with Artur Niemczewski, NED at CII
February 2017, insurance press is swarmed with predictions based on Carl Frey’s study that robots are coming and “topping the list of jobs at risk of being automated are insurance underwriters, with a 98.9% probability of being replaced by robots”. 6 years have passed and here we are, World Economic Forum marks Insurance underwriters as those who are “most likely to benefit from AI augmentation – 100% of their tasks have the potential to be augmented, the Jobs of Tomorrow report says”.
Together with my guest, Artur Niemczewski, Non-Executive Director at Chartered Insurance Institute, we discuss how the job of underwriter went from endangered to potentially 100% AI augmented in 6 years. And this time, we have a mystery guest as well…
Ewa Banaś: Welcome to another episode of IT Insights InsurTalk by Future Processing. In February 2017, the insurance press was filled with predictions that Insurance Underwriters faced a 98.9% probability of being replaced by robots. Fast forward six years, and the World Economic Forum now identifies underwriters as those most likely to benefit from AI augmentation, as 100% of their tasks are likely to be augmented according to the Jobs of Tomorrow report. Joining me today is Artur Niemczewski, a non-executive director at the Chartered Insurance Institute (CII). Artur, I have heard you are into extreme sports; is that true?
Artur Niemczewski: Yes, outside of insurance, my passion is kite surfing, which I have done for over 25 years. It is not exactly a corporate sport, so I was pleased to accept an invitation to the Lloyd’s Regatta, where I could focus on team building and learn from experts like Paul Blowers.
Ewa Banaś: How did you actually get involved in the insurance industry?
Artur Niemczewski: Most people stumble into it, but my background is as a trained engineer. I hold a PhD in nuclear engineering and a Master’s in Public Policy of Technology from MIT. People often ask what a “rocket scientist” is doing in insurance. I became passionate about it because it is a complex discipline requiring intellectual curiosity and the application of statistics and technology.
Ewa Banaś: Why are you so interested in technology, AI, and the underwriting area?
Artur Niemczewski: By trade, I was a broker rather than an underwriter, having served in leadership roles at Willis and Marsh. We always had a dialogue – and sometimes misunderstandings – with underwriters regarding technology. Fifteen to twenty years ago, I sat on the Market Reform Group (now the London Market Group) trying to implement electronic trading tools, but we faced constant baffles. I wanted to understand how we could use technology for better customer outcomes and more efficiency. At the CII, we have been interested in AI for a long time, publishing a digital code of ethics five years ago to address professional behaviour when digital tools take on decision-making.
Ewa Banaś: Today, we have a special guest joining the chat: ChatGPT. I asked it to respond as if it were in the world of The Matrix. ChatGPT, what did the life of an underwriter look like five years ago?
ChatGPT: The underwriter of yesteryear laboured in a cubicle maze, pouring over paperwork and spreadsheets in an endless loop of assessing risks. Today, the script has changed slightly as automation and AI have infiltrated the space, though the core tasks remain.
Artur Niemczewski: That description is quite accurate, complete with the flickering fluorescent lights. It is fascinating to see how the narrative shifted from extinction in 2017 to augmentation today. In the speciality risk market, underwriting still relies heavily on judgment and understanding which brokers to trust. For the last five years, we have seen an evolution from complex spreadsheets to macros and eventually Python code. However, these are often manual processes; if the person who wrote the code leaves the company, no one knows how it works.
Ewa Banaś: What is triggering the change today?
Artur Niemczewski: The Placing Platform Limited (PPL) was boosted by the lockdowns, pushing people toward electronic communication. The Next Generation version now transmits much more data. The problem now is that there are too many fragmented solutions. Every startup offers an AI-augmented rating engine, an underwriting workbench, or a data augmentation tool, but they are not integrated. Underwriters now spend their day figuring out which tool works rather than looking at spreadsheets.
Ewa Banaś: Is there a way to optimize this and make the underwriter’s life easier?
Artur Niemczewski: The vision—the Nirvana—would be a single dashboard where risks are submitted electronically and triaged. Red risks (outside appetite) are filtered out, green risks (within appetite) are priced automatically with the underwriter making the final call, and amber risks are where the underwriter adds real value through judgment and experience. For example, when underwriting 10,000 McDonald’s properties in the US, I want the data errors cleansed and the flood zones, crime ratings, and fire susceptibilities displayed on one dashboard via tools like Hazard Hub. This world does not quite exist yet because many past initiatives were point solutions that didn’t solve the whole problem.
Ewa Banaś: Do you think companies are investing enough to reach this ideal version?
Artur Niemczewski: Everyone is investing because it provides a competitive advantage; it is an arms race. However, at least half of that money is wasted on projects that exist for the sake of having a project. We have the building blocks, but we are missing system integration and APIs to put them together into a “house”.
Ewa Banaś: ChatGPT, what is your vision for the future of underwriters with AI augmentation?
ChatGPT: The future is a strange dance with machines. Underwriters will become Masters of the Matrix, using digital tools for analysis, but they must guard against obsolescence. They must adapt or face deletion.
Artur Niemczewski: It is a bit scary, but I agree with the sentiment. AI will not replace underwriters, but underwriters who use AI will replace those who do not.
Ewa Banaś: Thank you, Artur, for sharing your insights. Thank you to our audience for listening to IT Insights InsurTalk.