top of page

Rise and fall (and rise again) of AI


AI is being sold as the trillion dollar market opportunity. Everywhere you read, we see claims that it will be the one technology that will have the biggest impact on, not only the insurance industry, but the world.

Where are we now? It really depends on who you ask. Bleeding edge Insurtechs are raving about the potential of their ground breaking solutions, whilst the larger technology firms see the potential and are investing heavily but still feel it is early days and as for Insurers, they are largely playing a ‘wait and see’ game as they experiment with innovation labs and navigate the many operational and regulatory hurdles to bring these innovations to market.

Many obstructions still exist before this technology becomes widespread. To name just a few :

  • Data regulation

  • Ethical concerns

  • Technology maturity

  • Trust

  • Accuracy of training datasets

Let’s not forget about the AI Winters of the 70’s and 80’s where the technology hype quickly outpaced the actual progress made, leading to disillusionment and eventual funding pull back.


Its interesting looking at why these pull back’s happened. During the 1970’s much research was done into the basics of neural networks, an underlying technique still at the heart of AI research today, but of course back in the 70's computational power and training data was simply not available in anything like the abundance it is today, all of which hobbled the technology. The 80’s winter was different in that a huge amount of hype had been generated around the expert system paradigm often involving Lisp based languages, these systems could never get beyond a certain basic level of sophistication again failing to live up to the hype.

We are now in the midst of the next wave, computational power is prevalent, data is pervasive, and the actual technologies have been so commoditised that Data Scientist's can configure and start training a complex, deep, neural network in days using the cloud computing platforms.


This wave has legs!

The AI Industry

Some of the figures about startups in this wave are truly astonishing. 

For instance more than 3,000 start-ups, focusing on AI, were founded last year and almost more surprising, the top ten AI funding rounds brought in more than $8 billion dollars in total in 2019. Predominately, these companies tend to be based in the US or China with the majority starting within the last 5 years.

Companies By Funding Date.png
Companies By Country.png

Exhibit 1: AI Market Analysis

Source: InsurePro Research, Crunchbase

These companies can be split between those directly targeting specific industries:

  • Babylon , with healthcare

  • Lemonade, with Insurance

  • Argo with Self Driving Technology


and those that are building flexible technology stacks that can then be applied across industry boundaries:

  • UIPath, RPA engineering

  • Horizon Robotics,

  • MEGVII, with facial recognition

  • OpenAi, attempting to create Generalised Intelligence!

The foundations are now there, but there a long way to go before these tools can be used universally across the Insurance Value chain

The Insurance Market

To date, AI implementations have focused on distribution or point solutions improving processes or results for a specific use case, as capital, regulatory and compliance issues make it difficult to set up new end to end insurers. Indeed we may see the new MGA models kick-starting the more holistic uses of AI in the Industry.

We have shown below some examples of companies using AI technologies in the marketplace. What is clear is that the start-ups that are scaling are getting traction, signing partnerships and prototyping their solutions in the market. 

Every insurer needs to have the structures & staffing in place to take advantage of AI technologies, it is clear that within 5 years the benefits will be so great that this will be a hygiene factor for the industry.

The Insurance Industry needs to continue to work with regulators to ensure that as these technologies evolve, the scope exists to gain advantage from them.

Exhibit 2: AI in Insurance

Insuretech Examples.png
bottom of page