From Prediction to Transformation across Claims, Mike Daly

"While the popular view is that insights are the key benefit of artificial intelligence, in truth AI creates value by improving the quality of decisions. The good news is, the opportunities for it to do that in business are countless. But because decisions in one area of an organization usually have an impact on decisions in other areas, introducing AI often entails redesigning whole systems. In that way, AI is similar to groundbreaking technologies of the past, like electricity, which initially was used only narrowly but ultimately transformed manufacturing."

Ajay Agrawal, Joshua Gans & Avi Goldfarb in HBR Magazine Nov-Dec 2022 edition

Two decades after Edison switched on the light bulb, only 3% of U.S. businesses used electricity.

AI has also taken longer to adopt and deploy across claims operations and needs time for insurers to experiment and leverage it effectively. The frequency with which AI-powered applications appear on LinkedIn, insurer annual reports and websites, and trade publications you would think it is a mature, deployed technology. Read the case studies for automated damage and repair cost evaluation from technology partners such as CCC, Mitchell, LexisNexis, Verisk, Tractable, Solera, SLVRCLD, and Sprout.ai. and you might think straight-through-processing was “de rigueur”.

AI delivers predictions from large and multiple data sets; data is the raw material for AI to access, surface, normalize, analyse and predict outcomes. Digital maturity is nothing without the power of AI. Yet there is often a disappointment when trying to measure success and justify the expenditure. Why is that?

AI-powered applications are like unconnected digital islands in a systems ocean of data tides, waves, and shoals all flowing around the land masses of systems, unconnected technology stacks, and networks. Then, technology vendors add flexible, moving clouds as an alternative to fixed islands. A beautiful scenario when the sun is out, the seas calm, and the wind acceptable but dangerous when gales run amok, and storms batter the mainland and islands where much legacy technology is still located.

You need to be engaged and evolving with the market leaders to fully leverage these applications but you must avoid the problems of point solutions that do not connect and contribute with the wider network that is an insurer and its partners.

HBR analyses the problem in "From Prediction to Transformation" - why it happens, and the solutions.

“AI predictions often improve the quality of specific decisions but can have a disruptive effect on overall systems of decision-making. “
Decisions are usually a combination of prediction and judgment and when extremely accurate predictions are made by AI it speeds up the actual process to impact other people- ripples spread out from island to island, cloud to cloud.
The predictions hit reefs- This leads to a requirement for systems that balance modularity in design with co-coordination.
AI-powered applications are evolving but are like the island and floating clouds above. They require an environment that hosts them and delivers an optimal balance of modularity and coordination.

Systems platforms that connect the many islands, enabling modularity in design to advance that enables decisions to be made over a wider network.

RightIndem is one of the new breeds of cloud-native, API rich and micro-services architected platforms Its strategy was, and is, to be the claim technology partner to enable modularity in design with coordination to enable decisions to be made over a wider network. Automated and human decisions in the optimal combinations required.

Take a simple example: a customer registers FNOL via her phone and answers questions posed in a comforting bot framework thus automatically explaining her TV is broken; caused by her tripping over the dog, falling into the TV with iron in hand, smashing the screen before over-turning the whole TV causing further damage.

RightIndem automatically interrogates the policy admin and/or claims center core system in the background to validate coverage.

She added photos as automatically requested, five of them including the damage, serial number, make and model, and, very helpful, a video with an audio explanation showing the context of the whole event.

The automated RightIndem claims process confirms the claim is covered, analyses the metadata in photos and video, and “green” flags that the photos and video were taken in that location, at the time described and the images were not doctored. The claims software decides that the customer can be offered the replacement value all in an automated process (could all be rules-based or AI-powered).

But- here’s the thing. Automation has speeded up judgment and action, but will it hit an analog reef? Without the planning and strategy to avoid it, the decision is moved to someone who must now make a judgment. What is the current like-for-like replacement, where can it be procured, what is the price, and what are the deductibles? And say the customer says “cash will do fine thank you” we must now pass an instruction to the carrier’s payments system. If unlucky, the customer must wait for a cheque; lucky might be waiting days for an electronic transfer to her account.

OR- RightIndem enables the carrier to extend straight-through processing (STP) with point solutions like SLVRCLD, Be Valued, or Value-Checker to find the like-for-like replacement, cost, and availability. Then RightIndem applies any deductibles and offers in the same chat-bot style the choice of cash, a virtual credit card, or an Amazon voucher to redeem online. She presses the button of her choice, and the claim is settled.

RightIndem- a platform that connects all the islands, for which read modules or applications, passing data and decisions from one to the next and can even then instruct an automated payments system like Imburse or Stripe to pay by cash transfer, virtual credit card, or Amazon voucher then and there. The RightIndem strategy

That was a very simple example to be fair, but many claims are simple. Take a recent deployment of RightIndem for home and contents claims: -

Following a 10-week analysis of 760 home claims processed on the platform 87% of claims submitted were for single items only and therefore could have been processed via STP

33% of customers submitted claims outside of office hours and this is increasing.
80% of claims received had photos or videos attached.
35% of photos/videos received had referral indicators (metadata flags).
83% of customers who started the journey completed the digital journey.
66% faster settlement by handlers as they had more valid information allowing validation.
4 minutes was the fastest submission time – the average was 14 minutes.
91% of customers completed the CSAT.
82% of customers recommend the RightIndem platform.
86% said it was easy to use.

Hold on though- many property claims are more difficult than that. Internal damage from escape-of-water (EOW), multiple rooms damaged by flood and fire requiring emergency accommodation for clients and disaster/restoration attention. Some have vehicles involved and personal injury.

An eFNOL platform like RightIndem can apply triage logic leveraging custom question sets to drive the loss into the right parts of the organisation and its supply chain: e.g.

Desktop Survey
Liability@ EOW a water company responsibility or the insurers?
Loss Adjuster review
Builder/engineers, surveyor’s inspections
Disaster Restoration/CAT team

As you extend that into supply chains, the complexity of data mapping across all the participants in a claim becomes a challenge. BPO companies like Crawfords, Davies Group, Claims Consortium Group that deal with supply chains manage that with a mix of technology amd human intuition. When it comes to property supply chain management from main contractors to individual professionals you need the data modeling capabilities of an ecosystem platform like CoreLogic.

From a property perspective: Those buffers are smoothed out by creating a single data model across the lifecycle of claims

Looking at motor insurance, take the case of a road traffic accident involving the claimant’s vehicle and third parties.

Tractable, CCC, Mitchell, Solera, Audatex, and vendors can leverage AI, telematics, photos, and video inspections to estimate the parts required together with cost and time.When it comes to prperty

But this involves not just the carrier and the customer, but also the repair network, engineers, investigators interviewing third parties, and judging liability, possibly total loss. The number of hand-offs between systems and people increases geometrically and even a small 2% error rate at each hand-off would lead to significant data corruption and customer dissatisfaction.

The introduction of AI into a company’s decision-making doesn’t affect just you. It also affects your partners in the value chain and the ecosystem you operate in.

Or as HBR describes: -

What this shows is that while AIThis shows that while AI can , that effect doesn’t spread to decisions throughout a system. The fundamental problem—that demand needs to be aligned with supply—hasn’t really been solved. Like a stone thrown into a pond, your own AI solution has ripple effects on other decisions in the system.

That leaves us with something of a paradox. The value of AI comes from improving decisions by predicting what will happen with factors that might otherwise be uncertain. But a consequence is that your own decisions become less reliable for others. Introducing AI into the value chain means that your partners in it will have to coordinate a lot more to absorb that uncertainty.

To adopt AI “point solutions” means you must plan and coordinate systems to align effort and resources.

Furthermore, you need a platform that empowers you to combine coordination with modularity. Passes the decision from one AI application like Tractable to another like BAIL. That connects an investigations partner combining technology and human intuition into the whole subrogation process.

That is at the heart if the RightIndem strategy, technology and people that make up the company. We have found that working closely and in collaboration with new clients we can deliver a clams technology “stairway to heaven” as Geoffrey Moore of “Crossing the Chasm” fame describes.

This is a staircase of steps from the initial adoption of an AI package to help one part of the claims operation to the next step that, say, connects the automated estimation of damage and repairs/replacement/restoration to the skilled tradespeople, contractors, specialists, and engineers that constantly amend and update costs and times as supply chain disruption impacts the whole process.

The claims platform must enable coordination and communication and combine automation, human intuition, and knowledge.

It will not be simple; it will not be easy. But it will be feasible when a true partnership of joint commitment, preparation, and planning leads to the viable “stairway to heaven”.

To begin that journey, we would need to build clear representations of your current state and your desired future state. Capturing your current state involves an act of description.

Capturing your desired future state requires an act of design. If you do not have a clear design for your future state, you have no north star by which to navigate your digital transformation.

So, let us assume we have a clear design for our desired future state. It won’t take you long to realize there is little chance that a single intervention can get you from here to there. So, the next major deliverable must be a roadmap organized around a maturity model. Each step up the stairway should be designed to deliver value upon completion, thereby allowing the organization to pace its change management activity. Funding things as it goes, building its confidence and reassuring its various stakeholders.

With such a roadmap in place, now you have a current-state/future-state accountability mechanism that can govern each stage of the transformation—the software and systems, the systems integrators, the process owners inside your enterprise, and the people responsible for executing the processes.

There are many viable applications and data providers that leverage AI. machine learning, NLP, and RPA that an insurer can incorporate into their roadmap to harness prediction to help transform claims. When a carrier, broker, or MGA wishes to select the optimal solution for its specific needs it helps if it already has a claims platform that can host and benchmark the short-listed solutions. That's one test of a modern digital claims platform - being able to deploy production-ready trials.

And there are so many options to choose from - the list below is not exhaustive but shows the challenge facing claims, technology and supply chain leaders in carriers, brokers, MGAs and TPAs. From claims platforms to counter-fraud, to damage-estimation software to payments......

Ecosystem Claims Management Platforms

CoreLogic for property repair and restoration
Verisk for property and auto repair, restoration, total loss decisions

New Claimstech Management Platforms

360SiteView
Claims Genius
Claim Technology
RightIndem
Salesforce Industries (Insurance)
Snapsheet
Synergy Cloud

No one platform will have everything an insurer, broker, MGA requires. You will need to add third-party solutions to deliver all the requirements an insurer, broker, MGA will demand. Whilst all vendors claim to have large API libraries and integration capabilities many will lack the resources and commitment to be able to connect the required mix of third-party apps and data sources.

Point Solution Software

Policy Admin, Claims Validation, and Triage

HUGHUB
Iotatach
Open GI
Sprout.ai

Claims Damage and Cost Estimation (often combinations of these)

Be Valued
CCC
Claims Genius
ClickIns
LexisNexis
Mitchell
SLVRCD
Solera/Audatex
Symbility (CoreLogic)
Tractable
Upptec
Value Checker
Verisk
Xtract360

Property Repair & Restoration

Next Gear
Westhill

Liability Assessment

BAIL

Counter Fraud

360Retrieve
BAE NetReveal
Shift

Payments

Imburse
Mastercard
Stripe

Key Data Sources for claims management

Insurance was founded by leveraging the best of data to price risk, provide protection and manage business. “Data-driven” is a familiar refrain but the question is which data and intelligence is best and how can I integrate it into my systems?

Some platforms are positioned as key data providers e.g., CoreLogic for home and property from selling to protection and Verisk for auto and property. Their offerings are most complete in their home territory, North America, and expanding in Europe especially the UK and DACHS regions.

Claims platforms generate data in real-time and are a prime source of intelligence for decision making, fuelling AI, machine learning and rules engines including counter-fraud.

In the New Legacy platforms AI and ML is generally “on the side” in different database silos as the operational database was not designed to support real-time analytics. Being separate the learnings from the AI/ML cannot be connected to the core claims platform. The newer ClaimsTech vendors deem data, data science and analytics central to their platforms.

There is a frightening amount (circa 80% of unstructured data hidden in data silos, web forms, emails, SMS messages, voice files) that insurers must be able to access, normalise and analyse. Yet that data too often lies hidden and the value unrealised.

By the time the data is presented in dashboards, reporting or alerts it is often too late to take timely action and the costs of that failure are high.

New ClaimsTech platforms address that issue and whilst New Legacy claims platforms will you must beware of the potential complexity, time and cost involved in connecting all those silos and especially delivering real-time insights.

In addition, insurers need to license external data to feed the AI applications penetrating all parts of insurance with a selection listed below. Again, you will want to ensure that claims platforms and core technology can surface and leverage these.

CoreLogic- manage property data for selling, financing, and protecting property
Hazard Hub- property risk data
ICEYE- global flood earthquake and CAT damage data in near real-time
KETTLE- house-by-house risk assessment across USA
LexisNexis- vehicle, ADAS a,nd home data
McKenzie Intelligence Services Ltd wide range of data sources and data management
Mitchell- Auto data
SAFEHUB- building-specificS earthquake damage
Synectics identity, financial a,nd fraud data
Terrafirma property risk data
Verisk- auto and property data
WeatherNet- granular and near real-time weather data
WhenFresh- wia de range UK property data