Revolutionize Auto Insurance Underwriting with AI-Driven Risk Assessment

Auto Insurers Have a Long-Term Problem

The property and casualty (P&C) insurance segment has seen some very difficult times of late, dragged down in large part by the auto segment. While many point to the pandemic and the slow recovery from that once-in-a-lifetime event, the auto insurance industry has been underperforming for longer than that as the auto segment experienced has underwriting losses in 11 of the last 12 years.

That’s not to say that the pandemic and its aftermath haven’t played a role. When drivers got back into their cars as the economy began to recover, claims severity jumped due to high general inflation, supply chain issues that drove up parts costs, shortages of skilled repair labor, and accelerated litigation.

According to Fitch Ratings, full-year 2023 personal auto results finally showed improvement. While some of the larger carriers have returned to profitability, the 2022 baseline of an industry combined ratio (CR) of 104% wasn’t a great place to start. Smaller and less profitable carriers are expected to struggle above the 100% CR line for at least another year.

Searching for a Lasting Solution – AI Auto Insurance

One way out of unprofitability is to raise rates, which many insurers have done. US auto insurers raised rates by 11% in 2023, and 27.2% for the five years beginning in 2018 per figures compiled by S&P Global.

That short-term Band-Aid^TM approach can continue for only so long.  Insurers run the risk of pricing themselves out of the market and losing business to more nimble, more efficient competitors.

To make a sustainable change, auto insurance underwriting itself must be transformed.

The Benefits of AI Auto Insurance & Underwriting

Insurers are searching for the power of Intelligent Operations, or Intelligent InsurOps, which allows them to tailor every customer interaction to the individual. This provides the following benefits:

• Driving faster revenue realization

• Enabling market share gains

• Raising customer satisfaction

• Lowering costs

• Increasing profitability

AI auto insurance, including underwriting, is the key. Applying artificial intelligence (AI) and machine learning (ML) to auto underwriting leads to smarter underwriting decisions, and promises several key benefits:

• Increased speed – underwriting decisions and updates can happen significantly faster through the application of AI to the process

• Easy management of underwriting rules – business users (independent of  IT) can manage and change rules as often as they like, through table-driven and tree-based representations

• Rapid impact assessment – automated AI underwriting makes it easy to simulate the impact of risk and pricing changes on the portfolio before they’re put into market

• Governed deployment – underwriting management benefits from well-proscribed but rapid changes to reduce the risk of errors and recall of rules

• Improved personalization – AI-powered underwriting allows carriers to personalize offers to customers and prospects, increasing close rates and revenue

• Business agility – add all these factors up, and carriers see an increase in business agility that allows them to book more business, beat the competition, and improve profitability

Earnix is a leader in AI solutions for P&C insurers.

As a leader in mission-critical, cloud-based intelligent solutions across pricing, rating, underwriting and product personalization, Earnix delivers ultra-fast ROI that transforms how global insurers operate, by unlocking value across all facets of their businesses.

Its underwriting product, Underwrite-It^TM is designed to improve underwriting efficiency and results. It provides underwriting rules management, automated decisioning, advanced analytics, simulation, and business-driven deployment in a single package. It can be a key to improving underwriting for auto insurers.

3 Challenges and Rewards in Utilizing AI in Auto Insurance Underwriting

As with any transformational process, modernizing the auto insurance underwriting function consists of several steps and tasks, each with unique challenges but also with high leverage and consistent ROI when executed well.

In auto insurance underwriting, the challenges and rewards fall into three major areas:

• Data management

• Risk assessment and pricing

• Process improvement

Data Management

The vast amounts of data that insurers have at their disposal is both a blessing and a challenge. Sorting through all the data sources that they have access to requires intelligence and some amount of hard work. Data comes in all shapes and sizes and types, and must be dealt with through a variety of means.

Unstructured data

Unstructured data makes up the majority of what most legacy insurers have to work with. Analyzing this unstructured data was tedious work and prohibited insurers from reacting quickly to market changes.

Today’s unstructured data can encompass not just physical assets, but also digital sources, such as:

• Social media posts

• Emails

• Customer reviews

• Spreadsheets

• Images

• Competitive information

• Video and audio files

Even though the data exists in digital formats, these unstructured data sources are still disorganized and siloed. They are likely scattered across sales and marketing, pricing, rating, and underwriting internally, as well as with agents and producers in the field. Specialized software may be needed to bring some semblance of order to improve its usefulness in analysis (speech to text software, for example).

Structured Data

Structured data consists of discrete data points such as customer names, addresses, important dates, car makes and models, purchase data, payment transactions, and claims history. This data may be stored in:

• ERP systems

• Billing records

• Customer relationship management (CRM) systems

• Agency management systems (AMS)

• Other locations

Many insurers have supplemented these “traditional” structured data sources with behavioral data, such as is available through vehicle telematics that provide a wider view of driver behavior and to offer usage-based insurance (UBI) products. Other behavioral characteristics and purchase intent can be derived from website browsing and navigation data, providing clues from users’ search patterns and preferences.

Using AI in Auto Insurance Data Management

AI enables strategic and methodical management of an insurer's data assets to improve data quality and to make it ready for analysis and decision-making. 

AI enhances several data management functions, including:

• Gathering Big Data: AI systems can manage vast amounts of data from various sources, ensuring that relevant information is available for decision-making processes

• Data Cleaning and Organization: AI algorithms can clean and organize data, ensuring that the information used for analytics is accurate and up-to-date

• Reducing Data Noise: AI can separate important information from irrelevant data, eliminating noise, and helping insurers focus on valuable insights, saving time and money

• Handling Missing Data: AI methods such as imputation and predictive modeling can estimate missing values, resulting in more accurate—therefore useful—data, reducing the negative impact of missing data on studies and subsequent conclusions.

Data, of course, is a prerequisite to reaching the right conclusions, but to be useful all that data must be fed into the right predictive models.

Risk Assessment and Pricing

Insurance pricing strategy involves the interplay of risk assessment, actuarial analysis, market dynamics, regulatory compliance, and customer characteristics. Insurers need to strike a balance between adequately covering risks, remaining competitive, and satisfying regulatory requirements and fairness goals, all while striving for profitability.

The number of interrelated variables to be considered makes for a complex pricing process that involves all the functions of the business.

Applying AI-driven risk assessment and underwriting benefits auto insurers in the following ways:

• Risk Assessment

  Insurers must analyze the risks associated with insuring a particular automobile or fleet. Consideration involves such traditional factors as where the vehicle is garaged, its age and condition, the usage patterns of various drivers, and past claims history.

  Individual policyholders' characteristics, such as credit scores, claims history, and coverage options can also influence premium calculations. Insurers may offer discounts or surcharges based on these factors.

  More recently, through such technologies as Internet of Things (IoT) monitoring and the use of telematics data, driver behavior and usage have contributed additional data points that can be factored into risk assessment and have made the assessment more reflective of actual risk patterns.

  Earnix’s P&C expertise makes the analytics behind risk assessment and product formulation personalized and tailored to customers, regardless of their lifecycle stage, the auto(s) they wish to insure, or their competitive shopping habits. This expertise enhances carriers’ competitive position and ability to be a best-in-class provider.

• Underwriting

  Underwriting goes hand-in-hand with product and pricing. Over the years, and based on past experience, underwriting guidelines provide the criteria for accepting or rejecting risks and determining appropriate premiums. These guidelines may vary based on variables such as the insurer's risk appetite, financial goals, and regulatory constraints.

  Underwriting has traditionally suffered from a lack of automation and has been slow to catch up to other insurer business functions.

  While sometimes viewed as a staid function, carriers have begun to realize that underwriting can play a key role in business transformation, and have come to see the immense unlocked value in underwriting automation. They are beginning to act.

  Earnix Underwrite-It enables analytics-driven underwriting and empowers business users to create and manage decision logic, resulting in better underwriting decisions and increased revenue and profits.

Underwrite-It is part of a total solution for transforming insurance operations. The fully integrated suite improves traditional processes with AI-driven risk assessment, underwriting, competitive response, and personalization.

Streamlining the Underwriting Process

The traditional underwriting process is no longer suited to the speed, accuracy, and personalization needs of the modern insurer.

Deloitte outlines the characteristics of the traditional process, and its significant flaws:

• Manual processes: Understanding the current processes takes on a sort of “black box” aura, with too many hours spent in developing and deploying underwriting rules. Reuse is limited at best, meaning that new rules must be constructed repeatedly as business requirements change, and the process never seems to keep up with the need for timeliness.

• Disaggregation: Parts of the underwriting process are scattered across functional groups, including senior leadership, pricing, rating, regulatory, and IT, in addition to underwriting. Technology is similarly disjointed and difficult to manage and update, transparency and explainability suffer, and if the technology is in-house/legacy, processes slow down even further.

• Undifferentiated end-to-end solutions – tools and processes are quickly outdated, process flows do not have organizational buy-in across functions, and competitive advantage and competitive differentiation are difficult to come by.

It’s no wonder that by some estimates underwriters spend as much as 40% of their time on non-underwriting, mostly administrative tasks. This is a huge drain on underwriters’ time and expertise, which could be put to use for more productive purposes. And, it’s a significant drag on insurers’ bottom lines.

This inefficiency is not just an organizational problem. Underwriters feel the pain from an individual and professional viewpoint as well.

In a P&C Underwriting Survey, underwriters listed four key areas in which they would welcome increasing involvement, but have little or no time for, given they’re bogged down in administrivia:

• The development of new products and services;

• Process automation and the use of AI for improving underwriting performance;

• New data sources; and,

• ESG considerations

How do we get there? The answer is to adopt an automated, data-driven approach, and to implement centralized technology, stretching across functions including pricing, rating, underwriting, and product personalization.

The rewards are many and the ROI rapid:

Integrated workflows: Placing the entire underwriting workflows onto a single platform enables a single, collaborative process across functions. This eliminates the need for interdepartmental handoffs (and the attendant delays and opportunities for errors), significantly accelerates time-to-market, and enables internal teams to make the best decisions when it comes to risk, price, and profitability.

• Easier rules management: An integrated underwriting system allows the management of all rules in one place, enabling underwriting teams to update and deploy rules by themselves, reducing the burden on IT. Business users can update models, deploy changes, and manage once-complex logic through the use of table-driven and tree-based representations.

• The intellectual capital developed over the years need not be thrown away. By combining traditional rules-based underwriting with advanced ML and simulation, automated underwriting improves decision-making and simulation related to potential underwriting changes.

• Leveraged technology: AI and ML enable the development and adoption of new strategies, data insights, and a deeper understanding of customers and prospects. For example, AI algorithms can analyze vast amounts of data – from a wide range of sources – to identify previously undetected patterns and trends and better assess risk.

• Improved visibility into potential outcomes: Automated underwriting enables better-informed decisions faster and with less effort. AI insurance technology can provide a comprehensive, real-time view of risk profiles to help underwriters predict risk with much greater accuracy.

• Increased revenue and profitability: AI-driven analytical underwriting tees up the best decisions, leading to more appealing products and higher revenue and profits. Simulating the ways that various decisions will impact important KPIs and metrics, such as conversion, profitability, and retention, becomes easier than ever.

• Greater efficiency and agility: The combination of all these positive changes increases efficiency and underwriting effectiveness, delivers new business agility, and improves productivity, putting insurers in a position to gain a real competitive edge.

AI Use Cases for Auto Insurance Underwriting

There are a number of use cases for modernizing insurance underwriting using AI:

• Better understanding of risk: No matter how good the people and the systems, there is always a need to better understand and gauge risk. It’s one of those problems that doesn’t get “solved,” and automated AI-driven underwriting can contribute to a culture of continuous improvement.

• More productive data management: As outlined above, insurers’ vast amounts of data, especially unstructured data, are prime targets for AI to get under control and manage effectively, making the entire underwriting task more accurate and efficient. This helps ensure an environment of data-driven decision-making that everyone can rely on.

• More accurate and fair pricing: Automated underwriting can help ensure consistency in pricing for the same set of risks, as pricing can be generated from a centralized platform that cuts across functions. This reduces the tendency to generate differing pricing for common classes of customers and meets regulatory guidelines.

• Improved customer experience (CX): Throughout the customer lifecycle, beginning with initial contact and sales, automated underwriting can give the carrier a better chance of crafting a compelling offering, pricing it correctly, and closing the business. The request-to-quote cycle can nearly real time, turning it from days to weeks to nearly instantaneous.

• Once a customer is on board, AI underwriting can help tee up other offers with cross selling and upselling, and even detect when someone might be ready to leave, or churn, using sentiment analysis of customer interactions. Human underwriters can then step in to assist in the “save,” one of the higher-value tasks they want to be involved in.

• Supporting the drive for improved profitability: AI-powered insurance underwriting can contribute to improved profitability. Underwriters can deliver quotes with higher chances of conversion (improving top-line revenue), pricing in line with true risk (lower loss ratios), and optimal resource utilization (cost control). 

• Minimizing the chances for human error: Let’s face it, humans are subject to making mistakes, especially when addressing repetitive tasks and seeing the same issues repeatedly. AI and ML are indefatigable, no matter how many times they’re asked to do the same thing. Automated underwriting won’t remove humans from the loop, but rather frees them to pursue high-value tasks up, or to be the “tie breaker” for situations that require judgment.

AI-Powered Auto Insurance Underwriting at Its Best

Traditional insurance underwriting needs an overhaul, one that AI can vastly improve and accelerate. Carriers need to implement specialized underwriting software, and Earnix Underwrite-It is the answer.

Underwrite-It is based on composable, agile and real-time-capable technology, and is also tightly integrated with Earnix Price-It^TM and the Earnix Enterprise Rating Engine, drawing pricing, rating and underwriting together in a coordinated, intelligent layer of insurance solutions.

It’s time to move forward.