Takeaways:
- Insurers fail to provide enough coverage
- Regulation causes coverage gaps, but tech can solve them
- Buy vs. build choice will be key for insurers
Tanguy Catlin, senior partner at McKinsey & Company, is an expert in insurtech who follows the push and pull between carriers and tech providers. In 2023,
This article is from a longer interview and edited for clarity.
What are the insurance industry's biggest issues?
For hundreds of years, the insurance industry has been growing and creating value, expressed as the percentage of premium represented in the GDP. If you look at the last decades, all the growth in insurance, in mature market, has been coming from personal lines, 90 plus percent. And that growth is coming from increasing the price of insurance, not increasing the exposure.
We are failing society right now as an industry. We are not able to provide the coverage that is affordable to help our society benefit. There are three causes for that – the first one being that we are really complicated and provide a horrible experience when you look at NPS [net promoter score].
The second big issue is we are not providing the coverage that people need. If you look at all the deep catastrophes over the last 40 years, losses covered by the insurance industry are typically less than 50%.
Number three is we are the second worst industry when it comes to productivity gains over the last 30 years. On average, industries have improved productivity by 50%. Insurance has increased productivity by only 3%.
Now we have Gen AI, agentic AI, etc., and we think that those technologies can solve all those root causes. We can create personalized experiences. We can innovate products and coverages that are much more relevant. How we take advantage of technology as a catalyst for transformation that would bring insurance back into its mode of creating value for society?
What's causing these issues, especially the gaps in coverage?
Technology is not the primary cause of the coverage gap. But technology can be part of the solution to improve coverage. The primary driver is regulation. In California, they say you cannot increase price by more than 6%, and you cannot price in your predictions about future losses – when you are facing climate change. We're not surprised that the carriers are walking out of the market. That's why you have a coverage gap.
In car insurance, the level of technology that we have in cars today is unbelievable. But we've seen an increased incidence of death on the road despite cars being much safer. Why is that? It's distracted drivers.
But technology can be part of the solution to improve coverage. For example, if I can anticipate a pipe leak and intervene, I reduce water damage. If I can turn off access to your phone when driving, I will reduce distracted driving. If I can make insurance options more efficient, I can charge less and make insurance more affordable.
What's the next step that insurers can take to move forward on technology?
The answer varies based on your starting point and your investment capacity. In general, you will want to reduce your technical debt, make investments in data, and explore where to deploy agentic AI. Many will focus on driving significant productivity gains in the software development life cycle.
This will raise questions about architecture where you will need to decide where to buy vs. build while maintaining flexibility for future evolution of technology, protect yourself from vendor lock-ins and from your data being used to feed models that could benefit your competitors.
The answer will be very different for different actors. In general, you want to create an architecture for an agentic platform. You want that platform to be flexible enough to connect with solutions that you either build yourself or source from vendors.
We don't really know who the winners will be, so it's all about creating a flexible architecture, all open source. The second thing is that very few of those solutions are actually available. Therefore one of the questions you ask yourself is, are you willing to be a first adopter of technologies from vendors with two risks? Risk number one is that you lock yourself to someone else's computer architecture, and they might increase the price and hold you hostage. Remember, they're investing a lot of money, and they want to monetize it. Risk number two is giving them new data to train their models, that they then use to benefit other competitors.
Carriers want to move quickly, but need to decide whether they're going to buy or build. Building might be slower and more expensive at the beginning, but gives a lot more flexibility. Buying, as a first mover, has an advantage that outweighs some of the downward risks.
