Last year, Thomson Reuters completed a two-year programme designed to transform it from a content provider to a “content-driven” technology company. But CEO Steve Hasker said that, almost immediately after, “along came generative AI”, which he sees as again transformative for the group. 

Hasker, bets that the technology will transform rather than undermine its business of supplying information to lawyers, accountants, and other professionals.

Note that this includes all forms of AI as different tools provide different benefits and must be integrated effectively; data, analytics, AI, machine learning et al are a strategic subset of overall business strategy to deliver optimal competitive advantage and achieve company goals.

Extractive AI, for example, can handle an issue underwriters face. 

In the specialty and commercial markets one of the major issues is the sheer volume of email submissions (enquiries) they receive. The challenge is how to find the business they want to write hidden in the (hay)stack of emails and documents sitting in their inbox. One lead underwriter said that her team had to leave 90% of emails unread. And she wasn’t sure that the 10% they did quote was the best business! There aren’t many industries that have to ignore 90% of business opportunities!

To help solve this AI tools can review incoming emails and documents to help guide underwriters on what to focus on first. The underwriter remains in control, managing the data the AI is looking for, as well as deciding what to work on. The system should also learn as the underwriter works and responds, to better help identify the right type of business.  Training large language models on better-targeted and improved underwriting can create a ‘virtuous circle’ of continuous improvement and competitive advantage. but only by not ignoring the forgotten 90%. 

I attended an AI event hosted by Aimii at the House of Commons recently where blue-chip companies, the FCA,  and water utilities explained practical AI use cases. Many were to share corporate knowledge hidden in data silos and unstructured documents for the benefit of employees dealing with customer problems.  Turning enterprise data into answers

That is a typical issue facing insurers with 15, 25, and more incompatible technology stacks. One global carrier experimenting with AI in all its forms nevertheless is hobbled by 40% of its insurance data being processed on legacy AS400 systems. No wonder it still spends so much time manually analysing data from multiple data stacks to inform management of work-in-progress status.

The Aimii customers leverage the hidden unstructured data across these multiple data stacks using Aimii's AI platform supported by many data scientists and, vitally, data engineers to tackle these perennial problems.

They can analyse common causes of faults and damage in products from vehicles to aero engines, to nuclear power plants to water leaks thereby speeding up problem analysis and resolution. A simple solution with significant performance improvement. This parallels the use of AI by Thomson Reuters to deliver accurate, faster insights to lawyers, accountants, and other professionals.

That answers the constant challenge faced by underwriters and claims staff; the effort required to find the right information hidden within complex and unstructured documents when assessing a risk or claim.  

AI tools incorporated into digital claims FNOL help triage claims into the right workflows and teams. Simple accidental damage can be validated against cover and evidence to automatically settle these high-volume low -cost claims freeing up time for claims handlers to focus on the complex claims and customers who need support.

A lesson to be learnt from Thomson Reuters though is that these projects should not be started in isolation. Claims and underwriting are so intertwined that transformation and the leveraging of AI should be tackled in parallel and not separately.

Insurance has long suffered from heavy reliance on multiple documents across the business in all sorts of formats and layouts. Even something as simple as an inception date can have many different formats and titles. Traditional OCR (Optical Character Recognition) would get confused by that variation, needing much more consistency on where to find the expected information – even skipping to the next page would cause a problem. Newer AI-based data extraction technology for several years tackles a variety of document types. The introduction of the new Large Language Models (LLM)has meant a major improvement in the ability to return insights from messy data stores.

The other end of the spectrum covers the MGAs and brokers seeking capacity. Unless the business is to be one-sided, they need the earnings patterns, triangles, and predicted ultimate loss ratios to negotiate capacity and plan distribution effectively. They need the granular analysis of customers by geography, product, and claims, to plan for improved performance and growth and be ahead of the underwriter armed with these new AI-powered tools. Too many still rely on spreadsheets which are notoriously prone to corrupted links and inaacurate outputs. 

Predictive analytics company specialises in solving these issues for MGAS and brokers. 

AI and machine learning (ML) are still tools to be leveraged by the business in the context of the company vision, goals, and strategy. Insurers need the talent, capacity, and time to incorporate them into the management of business.  That is why I asked if insurers should take a leaf out of the book of Thomson Reuters. The sheer resources they are allocating to leverage these technologies are awesome. That includes acquiring AI companies, developing technology in-house, and, no doubt, seeking advice from systems integrators, consultants, and technology partners.

That is another lesson to be learnt. Far too frequently an insurer commissions a platform or software vendor to deliver a POC with carefully measured results demonstrating success  only to then decide to build inhouse with the help of incumbent vendors who are confident they can replicate the successful POC. All too often the new project gets put on the back burner of IT projects never to emerge. A state I have heard be named the 'Pit of despair'. 

The speed with which Thomson Reuters has pivoted strategically to be able to levarage AI is a warning that the approach above is dangerous. Such is the speed of advance in generative AI and LLMs spurred that using the same partners and legacy technologies is not a viable approach.  There are partners to help experiment with GenAI and LLMs and find the sweet spot between these and more traditional AI eg extractive and conversational AI. Platform providers that are incorporating these tools into their products whether they be counter fraud, underwriting workbenches, claims platforms with eFNOL, and straight-though-processing #STP and many more.

They all have the potential to free up scarce time and resources from mundane control and admin tasks to spend more time on performance improvement and the ultimate objective: - growth with profit. 

BUT- to achieve success over the whole value-chain of insurance from Selling and Quote & Bind through distribtion to settling claims is a complex matter. Not helped by the dependence on legacy technology surrounded by layers of new technologies in a way well visualised by this picture.


It's not just carriers that suffer from this vulnerable dependence on undocumented software written ages ago bu a developer who has long since left the business. 

It is important to look beyond AI as just another digital plaster placed over old legacy systems and processes. Underwriters will jump at any chance not to have to use the clunky policy admin system they’ve been required to use up until now. However, there are only so many layers we can put over these old systems. Too many layers make future change costly and slower to deliver. We need to be smarter with how we use these new tools to make a difference. 

That is why many insurers have taken the plunge to invest in modern MACH core platforms that empower insurers to be adapatable.  It's not an easy decision and requires the level of commitment that Thomson Reuters applies with vision, planning, great leadership, buy-in across the business and commitment to success. 

From large enterprises like carriers Liberty Mutual and esure partnering with EIS, to nimble MGAS like Arma Karma partnering with Genasys. Instanda, ICE, Ignite, Novidea, HugHub- all these modern technology platforms are helping insurers breaking free of the chains that legacy systems imprison and burden them with. 

The AI-powered technologies will help improve efficiency and performance. Companies have leveraged AI to make process mining much more efficient and effective. With the agreement of staff ( a vital factor) , new intelligent monitors can watch a person work across different systems and steps, looking for where people get stuck or waste time looking up information or fixing data errors. These new AI-driven tools don’t just map the process but build the workflows to help staff overcome those delays and accelerate the route to greater efficiency. 

Help is the keyword as insurance is bound by regulation and transformation must proceed in step with compliance, security, ethics and customer acceptance. Insurers have culture in their DNA and technology adaptability requires cultural adaptability based on trust in the company leadership and buy-in of the company vision.  AI needs to be managed by humans as much as it helps humans manage. 

Can AI help map out the future and the strategies to anticipate them? AI can be used as a sounding board to test thinking, speed up research, and act as a counterweight to internal bias and inward thinking.  Caveats should be heeded, however. Fact check and use your own experience and knowledge to sanity-check answers.The creative process can be helped along by using the right AI tools with the right data sources but not replaced by AI. Chief Strategy Officers, CEOs, Chief Transformation Officers, and teams working on scenario planning. This is where critical thinking, lateral thinking, analytical prowess, and asking the right questions are essential.   

Asking strategic questions, such as “What can I do to help myself?”, “How else can I do this?”, or “Is there a way to do this even better?” 

While some people doggedly work toward a goal or respond to a challenge with a gut feeling, others are predisposed to a strategic mindset when faced with a task. Those who display a strategic mindset constantly review what they are doing and their methods. They actively analyze the task, plan, self-monitor, and revise their strategies.  

AI can help by surfacing all the data available and presenting relevant facts to those planning for a changeable future.  A strategic mindset helps identify today's outliers that may be tomorrow's future. Generative AI is less useful here as it predicts from the past. Expert practitioners can help promote and train it to be more focused on significant outliers indicating change and disruption. But it needs a strategic mindset to make the leap, be bold, and take quantifiable risks.   

Who bet on Tesla outstripping the value of traditional internal combustion-engined vehicle manufacturers in its early faltering stages? Or Space X and re-usable rockets paving the way to mine rare metals and minerals from far-flung asteroids to sustain electric and autonomous vehicle and solar-powered energy when early prototypes failed? 

Insurance may seem a little staid and traditional at times but its future will be ever much as exciting and different and to anticipate that AI and machine learning in its various guises is an essential part of the future growth jigsaw.  

Thomson Reuters is committing massive human and financial resources and a strategic mindset to invest in and leverage AI, machine learning, and, no doubt, quantum computing. Insurers need that resolve and to partner with the companies that will help apply that level of strategic mindset.

Fancy a chat to talk through these issues?  Drop me a line to about underwriting and AI, for claims transformation or or let's meet up at Insurtech Insights in London March 20th-21st. 


Further Reading

Don’t Let Gen AI Limit Your Team’s Creativity  HBR March-April 2024

Elon Musk uses a simple strategy to reach his goals—here it is in B2 2021

Exploring the impact of AI on the insurance industry: benefits, challenges, and future directions in Trading Herald Feb 2024

Helping insurers cope with general and claims inflation, rising premiums, surging complaints, and adverse loss ratios my guest blog for Genasys

AI code of conduct for the insurance claims and supply chain industry launched Insurtech World

Saints & Sinners - proving effective AI use cases in insurance Insurtech World