The sheer volume of hype about GenerativeAI, LLMs, ChatGPT for industry-wide transformation and smaller projects is overwhelming. Saints will offer advice based on actual practice together with all the caveats that must be addressed. Sinners rush in where angels fear to tread. But of one thing we should be certain- you should be experimenting now and know why while focusing on achievable first steps.

"Should we automate away all the jobs, including the fulfilling ones? Should we develop non-human minds that might eventually outnumber, outsmart...and replace us? Should we risk loss of control of our civilisation?” These questions were asked back in March in an open letter from the Future of Life Institute, an ngo. It called for a six-month “pause” in the creation of the most advanced forms of artificial intelligence (ai), and was signed by tech luminaries including Elon Musk. It is the most prominent example yet of how rapid progress in ai has sparked anxiety about the potential dangers of the technology."

The Economist

On the other hand, you might be amazed at the innovation being implemented by a multitude of insurers and technology partners from underwriting and pricing to claims management, counter-fraud to supply chain management.

Large insurers could benefit from savings north of £100m by developing several simple use cases, or improvements to their combined ratio of 5-6%

'A working paper from earlier this year, by Erik Brynjolfsson of Stanford University and co-authors, studied the effect of equipping contact-centre agents with an ai-based conversation assistant that offered real-time suggestions for responses. The agents remained in control of the conversation, and were able to accept or ignore the AI’s suggestions as they saw fit. The authors found that the tool increased worker productivity by 14%, measured by the number of chats agents were able to successfully resolve per hour. It also disproportionately improved the productivity of less experienced agents, meaning a more consistent service for customers.'

The Economist Sep 28th 2023

Look- there are many dangers inherent when implementing LLMs, Generative AI, ChatGPT, et al and you need vendors and consultants that know this and protect insurers from these. and also advise when other AI technologies are better suited to achieve the goals you want.

So I was impressed to read of this initiative by Palantir which has proven its capabilities to governments, military customers, banks, healthcare, and industrial giants like Airbus which has transformed its complex supply chain. ( Please note that I have no commercial or other involvement with Palantir).

Bring a team of data engineers/scientists, business analysts, LOB leaders to a local boot camp lasting from one to five days, a desirable use case to work on, get those keyboards and minds working, and have a working MVP by the end to carry back to your company. Yes in the UK as well.  What's not to like?

Apply AI in your business environment assisted by experts from Palantir who have achieved success across so many industries. Get started faster, remove friction, build in your intuition, iterate and start proving the value of your use case.

Surface and analyse the unstructured data required to achieve your goal, connect the many business processes outside the core system ( plus those inside of course).  Build an MVP that proves the potential so you can get the funding, support, and resources to build out a production-quality product/service.

Palantir obviously hopes that these bootcamps will prove their capabilities and win them new customers. Nothing wrong with that and they are putting their skin in the game resourcing and supporting these bootcamps. Thereafter it is up to you. Take it or leave it. Go back to your existing technology partners but do ask them and yourselves a very important question.

"Why haven't we proved and deployed these use cases already?

So many times an insurer has enlisted one or more technology providers to deliver a proof of concept with a goal of achieving measurable KPIs and on a successful conclusion, sometimes exceeding the original goals, only to be stalled by the incumbent vendor saying "We'll do that in 12 months" Too often they never do being over-stretched completing current projects, upgrades and new versions.

The project leaders often wilt under pressure from the C-Suite to focus on the next quarter's Combined Ratio and Central IT to use what they already have instead of making the business case strongly for innovation. 

So back to leveraging AI to tackle the Experience/Efficiency Paradox a term coined by Rory Yates CSO at EIS from whom I quote below.

" Look at the transformation of claims handling. The work is traditionally seen as isolated from the business, and the transformation is conceived and run by a team. But, the efficiencies that can be gained from a data-driven model that's centered on the customer can get missed when addressing claims efficiencies in isolation. For example, when all customer data is up to date and automatically integrated into the claims experience, it becomes a far more intelligent capability.

The data integration immediately creates a greater focus on how the experience can benefit the customer. Policy details and changes in circumstances appear in real-time when the claim is raised. There’s no need for policy changes to be a “change request” for the claims team to act on. Manifesting a product change is much more efficient in every customer interaction or experience throughout the business.

This works both ways. Consider a protection business that ensures changes in its customers’ lives are tracked and acted upon. Those updates can highlight a need to address changes in coverage, while also delivering sought-after efficiencies. For example, a customer moving to a new house creates an opportunity to reassess their coverage, build a deeper relationship, and demonstrate a keenness to be of service.

Rather than running data-driven communications as a project or change request, you have a continuous and adapting relationship. This relationship makes sure every change in a customer's life is understood in the product context. 

From using open banking or employee records to identify changes in financial circumstances to reacting to a change of address, these aren't only new opportunities to create vastly better experiences, but they can also drive huge efficiencies. 

Automating data integration is in stark opposition to current customer experiences. A colleague recently told me they had three policies with a provider who couldn’t see them collectively and recognize that relationship. This inability made a simple address change a triplicate exercise.

Putting the insurer by the side of the customer, and completely changing the paradigm for the product and how the business works, can turn large and often costly business processes into continuous and seamless customer experience outcomes".

The Experience/Efficiency Paradox by Rory Yates published in Insurance Thought Leadership

Palantir has given insurers the opportunity to go better than a POC and deliver a minimal viable product (MVP) in one to five days. 

You can sign-up for a bootcamp here. Or if you want to discuss it before doing so drop Scott Thomson a message  and/or have a look at this short video

Note: I am not employed by, working in any capacity for, or receiving any remuneration for this article but offering it in good faith as good practice and a beneficial initiative for insurers. 

Further Reading

Revolutionising Pricing Using AI and Data Overlay

Where to use Extractive AI vs. Generative AI for enterprise 

UK car and home insurance complaints accelerate; is transformation working?

The data dividend: Fueling generative AI

The Experience/Efficiency Paradox

AI's value will come when we know it can do work for us

The Musk Disruption Doctrine