AI- all kinds and not just LLMs and GenerativeAI featured across the Insurtechs Insights conference in London last week and a common theme was that of adoption speeding up, experimentation with GenAI evident but that the latter needed human oversight to avoid some potentially serious dangers. It's not just hallucinations but completely made-up content presented as authoritative reference sources.

Amara's Law, named after the futurist Roy Amara, states that we tend to overestimate the impact of technologies in the short run and underestimate them in the long run.  So this article is not pessimistic; rather it points out that we should not be deterred just because it will take longer than you think to reap the fruits of AI and require the full support of the company to plan, experiment, find ideal use cases, prove them and operationalise AI. No short-cuts!  That is why an ideal use case in healthcare provides good insights into how insurers should leverage AI tools.

Diagnostics is one area that should be ideal. The speed of analysis, ability to train on large data sets, and structured way a diagnosis is approached using data, images, and symptoms to decide on prognosis.

Yet things are not so simple.

"The technology seemed so good that it would soon replace humans. But the reality has been rather different. Beyond a few applications — for example, in mammography and colonoscopy — evidence of AI’s superiority over doctors remains thin. Human intervention is still essential.

AI has the potential to significantly transform employment in many aspects of healthcare — from basic research to clinical care, patient monitoring and medical administration. But Wachter’s (a director of the Josiah Macy Foundation)  comments are a reminder that its impact in transforming human roles and the timescales for such changes to take effect remain far from certain. “I’m most impressed by our ignorance on AI,” says Wachter. “It has speeded up mammography, but there are lots of areas where it is not quite as good. Diagnosis is not just looking at digital dots but understanding and placing them in the clinical context. The number of scans is growing faster than AI is advancing.”

Claims processing, underwriting and pricing are areas where there are great hopes for the application of AI. Cytora and Hyperexperiantial, amongst others, are data-driven and AI-powered to augment underwriting and pricing teams. Keep an eye on those two.

The best AI tools are still dependent on the quality, relevance, and timeliness of data. Giroux.ai works with MGAs and Brokers to help them understand their data- the what, who, when, how, and where that helps them measure the health of the business and how to improve performance. That helps establish which insights are needed to make the best decisions for profitability and growth. 

Too many companies want to leapfrog the vital disciplines of data preparation, processing, and management before benefitting from predictive analytics, automation, straight-though-processing, and freeing up professionals to manage rather than be submerged in admin. 

I have heard of underwriters unable to cope with incoming capacity requests by email and ignoring 90% of potential business. Underwriting workbenches need to be able to ingest this unstructured data, analyse, and triage it so that underwriters don't miss 90% of potential business.CGI demonstrated the latest version of its underwriting workbench at Insurtech Insighs and how it tackles these and other challenges. 

Aimii a UK company is engaged by many blue-chip enterprises to help understand and leverage their data.

"On average, only 1% of your organisation’s data can be seen by employees. Gartner research shows 44% of employees have made a wrong decision because they were unaware of information that could have helped.

The Aiimi Insight Engine discovers and interconnects all information across the enterprise, so you can break down these harmful data siloes and safely put information into the right hands, faster."

Aimii is also helping companies experiment with and leverage LLMs and GenAI while not ignoring conversational, extractive, and other AI tools. 

The article on AI in healthcare has many lessons for insurers from Tier One carriers over personal lines, commercial and life, and smaller Tier Three carriers, MGAs, and Brokers. To quote the National Lottery you have to be “In it to win it” but without the right advice, the right tools, and the right data it may well have the same chances of success as the lottery.