Transforming Insurance Underwriting with Google Cloud AI— Part 2
We discuss key challenges across the building blocks of insurance underwriting and how Google’s AI and analytics rises up to the challenge
In the first post of this series, we discussed the key building blocks that form the foundation for striking an optimal balance between fully automated, low touch and full touch insurance underwriting.
In this post, we discuss the limitations of the building blocks impacting effective underwriting and we begin covering some of Google’s AI-driven solutions to address each of then. As a quick recap, in the first post, we discussed the following as the key building blocks for achieving underwriting excellence.
- Portfolio Steering
- Profitable Pricing
- Risk Selection
- Capacity Optimization
- Rationalized Coverage
- Submission to Quote Efficiency
Let’s talk about the key challenges insurers in general, and underwriters in particular, have to contend with, across each of these building blocks.
Submission to Quote Efficiency
Most often, insurers are bogged with manual non-streamlined workflows that tend to be a continued legacy of dated technology capabilities piled on over the previous decade, that result in everything from process delays to inefficient data intake on the side of producers or direct to consumer insurers. This results in slow, inconsistent underwriting — which becomes a fundamental bottleneck to let the underwriters make the best of what they already have.
Rationalized Coverage
Assessing the adequacy of coverage breadth has been another challenge for underwriters who are often stuck between the not whole to too narrow. There are also no efficient ways to track coverage nouns with quantitative metrics which would also be a key factor towards improving efficiency of a future claims management. As a side note, many insurers also fall into the convenience trap of keeping underwriting and claims as entirely separate functions, rendering both these critical and highly dependent functional areas incapable of leveraging the data from each other’s data sources.
As a result, insurers not only end up with an adverse reserve development, but have also opened up to a significant threat of claims leakage. A large insurer had recently also reported in their earnings that adverse reserve development remains a key problem for them and they have invested over $xxB in reinsurance to cover their own risk of adverse reserve.
Profitable Pricing & Capacity Optimization
Not having a unified view of data across key insurance functions like claims and underwriting also results in inefficient data mining over historic price movements. This results in inconsistent pricing for similar risk exposures eventually leading to an inefficient capacity assessment which in turn leads to significant policy losses.
It’s a cascading game that rolls losses from one level of inefficiency into an exponential loss on the next level, and so on. Insurers then have to contend with a reduced policyholder surplus, that results in a direct impact to their bottom-line.
Portfolio Steering
Not being able to tap all the data is also a factor in an inability to have adequate portfolio insight for assessing exposure management and forward-looking judgments on exposure variations. When insurers are forced to wrestle the proverbial data-octopus and are not able to connect the dots across their disparate data, they end up with a very narrow analysis of propensity of risk changes at the time of underwriting.
Risk Selection
Risk selection becomes more subjective than driven purely by data parameters. This also means not considering geolocation for risk assessment, spending a long time collecting or waiting for the right combination of data to be available. While some insurers now acknowledge the importance of geolocation-driven analytics of their data, obtaining reliable geo information that isalso constant refreshed continues to remain a major challenge.
If it’s not obvious already — the key governing force to cover all of the challenges above is to have these three capabilities.
- Quick ingestion of data for underwriters, so they have what they need, when they need it.
- Enable underwriters to connect the dots across all of that data, including additional parameters like geolocation.
- Provide the underwriter with a mature AI and analytics environment that can allow them to accomplish all the six building blocks we discussed.
Here’s quick snapshot of how capabilities driven by Google’s AI and Analytics platforms feature solutions to the building blocks.
Submission to Quote Efficiency
Adopting accelerated document processing to intake producer submissions at scale, process the ingested data at scale, transform them to suit your unique business needs, make it instantly available to underwriters — and doing all of this in real time.
Rationalized Coverage
Utilizing Natural language understanding to map language to policy metrics that would enable underwriters to quantify risk exposure, covering petabytes of data when necessary.
Profitable Pricing & Capacity Optimization
Unify data and capabilities for granular analytics for effective pricing and portfolio steering. For instance, doing a time series analysis of historical pricing changes.
Portfolio Steering
Creating an AI-Powered Data Warehouse and Data Science platform specifically for Underwriting. Expand analytics capabilities to cover internal and external data sources to understand impact of exposure changes to portfolio.
Risk Selection
Making Geolocation data as a key data source to enhance exposure management. Mature risk selection would involve a combination of several capabilities we discussed including real time data ingestion at scale, natural language understanding, analytics, geolocation mapping and entity resolution.
In this post, we started going deeper into key challenges and solution areas where Google’s AI and Analytics technologies can help insurers achieve underwriting excellence.
In the next post , we’ll discuss Google AI and Analytics technologies that can help transform insurance underwriting across each of the key building blocks.
