March 8, 2021 Auto Physical Damage

The Evolution of Claims Automation

<p>COVID-19 has hastened the pace of digital transformation in many industries, including auto insurance. When the pandemic began, carriers and collision repairers were forced to find creative ways to limit unnecessary in-person interaction between their employees and customers. This led to the <a href="https://www.mpower.mitchell.com/2020-auto-insurance-trends/">rapid adoption of virtual, or photo-based, estimating</a>.</p>

COVID-19 has hastened the pace of digital transformation in many industries, including auto insurance. When the pandemic began, carriers and collision repairers were forced to find creative ways to limit unnecessary in-person interaction between their employees and customers. This led to the rapid adoption of virtual, or photo-based, estimating. For insurers and repairers, virtualization supports the need for social distancing, accelerates the claims handling process, improves efficiency and increases customer satisfaction. According to LexisNexis Risk Solutions, 95% of auto insurance carriers are embracing virtual claims handling, with many setting their sights on the Holy Grail of automation: touchless estimates. By 2025, the global data and analytics company predicts that more than 80% of claims processed will be virtual, and up to half of non-injury claims fully automated. So what does this mean for the auto insurance sector? While the pandemic has illustrated the benefits of virtualization, truly automated or “touchless” estimates remain the ultimate goal. Achieving that goal, however, requires a technological evolution. Despite business incentives, customer expectations and external motivators like a global pandemic acting as catalysts, the transition from onsite inspections to full automation will take time.

The idea that, today, fully automated appraisals without human intervention can be accurately produced is fiction.

A Deceptive View of Automation

To understand the trajectory of claims automation, we must first address misperceptions about the current state of touchless estimates. The idea that, today, fully automated appraisals without human intervention can be accurately produced is fiction. In many ways, it’s analogous to recent exaggerated assertions about the prowess of self-driving vehicles. While technology advancements have led to increased automation, the need for human oversight and intervention remains necessary. Furthermore, the infrastructure and regulation required to support a completely automated end-to-end system are still evolving.

Why Touchless Estimates Are the Holy Grail

To stay competitive and meet increasing consumer demand for better response times and self-service capabilities, insurance carriers are on a quest for the Holy Grail: touchless estimates. Why? An automated appraisal process is expected to deliver:

For it to be successful, though, carriers must consider their own unique workflow requirements when introducing automation. A one size fits all approach may work for field estimating, but a scalable, automated appraisal solution requires an open, flexible and cloud-based ecosystem that can easily integrate with other critical business applications. Open ecosystems, like the one powering Mitchell Intelligent Estimating, give insurers the ability to streamline operations and leverage best-in-class technologies that continually reduce the reliance on human effort as they build out an automated claims experience.

Trends Driving Higher Efficiency

The evolution from onsite inspections to touchless claims is marked by three trends, or levels of automation. These levels are not sequential. Instead, they are being developed and deployed in parallel as the artificial intelligence (AI) used to automate estimates becomes more mature.

Virtual Estimating

Virtual estimating dramatically increases efficiency and demonstrates the power of using photo capture to produce accurate assessments. As mentioned, this method of inspection relies on images and videos. Although virtual estimating puts photos at the center of the appraiser’s workflow, it’s primarily human and not machine driven.

  • Customers start the claims process on their mobile devices, and are guided through how to capture and share photos and video of the damage.
  • AI is leveraged to categorize and organize the imported images as well as to detect and highlight sections of the vehicle with damage.
  • An application essentially provides appraisers with a 360 degree view of the damage, allowing them to write an initial estimate as if they were at the vehicle instead of at a remote location.

Prior to COVID-19, virtual estimating was used for low severity claims. Compared to other inspection methods, it accounted for just 6% of all estimates written in the United States last year and 4% in Canada. Following the start of the pandemic, those percentages quickly jumped to 13%. Field appraisers performing onsite estimates can typically complete three to four per day when factoring in administrative tasks and travel time. Working from images, however, allows the appraiser to finish approximately 15 to 20 estimates per day.

Guided Estimating

The next step in the progression from handwritten estimates to machine-written estimates involves human-machine collaboration. AI guides the appraiser through each decision point. The goal of guided estimating is to put more power in the appraiser’s hands while leveraging the AI for useful recommendations. At this level of automation, the machine is becoming much more involved in the process. Although the appraisers remain in control, the information and decisions presented to them are powered by AI with each sequential line of the estimate being suggested for the appraiser’s consideration. Guided estimating goes beyond virtual estimating by:

  • Driving a Set of AI Predictions That Recognize the Damage to Components
  • Turning These Predictions Into Repair Line Recommendations for the Appraiser
  • Surfacing Supporting Evidence and Empowering the Appraiser to Edit and Incorporate His or Her Own Expertise
  • Supporting a Continuous Feedback Loop That Allows Appraisers’ Decisions to Educate the AI

Automated Estimating

The final level of automation is, of course, fully automated or touchless estimating—including all operations, parts selections and pricing generated by the AI using vehicle and claims data. Here’s how this predominantly machine-driven process works:

  • Like virtual and guided estimating, touchless estimates start with capturing claim details and vehicle content.
  • This content is then analyzed using computer vision and other machine-learning algorithms to translate into component-level estimate lines.
  • The AI then pre-populates the entire estimate for the appraiser to review and approve.
By 2025, at least 40% of total loss claims will be settled within a couple of days instead of weeks.

Beyond estimate automation, additional data integration will help carriers further streamline the claims process. Ultimately, claims have the potential to be settled even faster by incorporating telematics incident reports, loan payoff amounts, titled/registered owner information, and taxes and fees into the workflow. In fact, Bill Brower of LexisNexis Risk Solutions predicts that "by 2025 at least 40% of total loss claims will be settled within a couple of days instead of weeks."

Partnering to Deliver the Best Outcome

The transformation from field appraisals to touchless claims isn’t done in a vacuum or entirely by a machine, for that matter. Appraisers are key to developing and improving the process. By introducing automation gradually, they can build their confidence in the AI while simultaneously perfecting the results through their ongoing feedback. Creating an experience where guidance is clear, actionable and, most importantly, transparent helps build trust between humans and machines. Processes driven by human-machine collaboration can lead to vast improvements in speed and accuracy, but they take time. That doesn’t mean that carriers shouldn’t get started today—they should. When it comes to meeting consumer needs and staying competitive, the question isn’t whether or not to automate appraisals. It’s which partner has the experience and expertise to help you achieve your goals by best supporting the evolution of your organization’s claims process.

Olivier Baudoux

Sr. VP of Global Product Strategy and Artificial Intelligence, Auto Physical Damage

Olivier Baudoux is senior vice president of global product strategy and artificial intelligence for Mitchell’s Auto Physical Damage division. A highly regarded technical leader and expert in artificial intelligence and automation, he excels in formulating technology strategies to meet diverse customer needs and managing development across global teams.

NEWSLETTER

Sign up to Learn More