Artificial Intelligence: Ready to Simplify the Auto Claims Workflow
The P&C industry has been buzzing for a while about the emergence of “smart” technology entering the claims workflow—from natural language processing to machine learning and artificial intelligence, the frequent tech talk has left us all bracing for the new paradigm. But with all of this discussion about technology getting smarter, how do we ensure that we are “smart” about how we use it? Exactly how and where do we first implement technologies like artificial intelligence to improve day-to-day claims workflow and decision-making? While artificial intelligence promises to be transformational over the long-term, A.I. must first gain traction by making tangible improvements that expedite and simplify the auto claims workflow.
As with any new technology, however, A.I. adoption as a standard part of the claims workflow will only reach critical mass when implemented with usability and practicality.To know where to apply A.I. first, we have to remind ourselves of the immediate problems that A.I. is best-suited to remediate. Fortunately or unfortunately, there is no shortage of opportunity. As Ryan Mandell, Director of Performance Consulting for Mitchell Auto Physical Damage Solutions explains, “With rapidly changing conditions that put more drivers and more complex cars on the road, it’s no surprise that auto claim volume and loss costs have increased substantially in recent years.” This naturally creates a challenge for carriers to improve claims outcomes while simultaneously absorbing a heavier workload and maintaining estimate accuracy and repair quality. Artificial intelligence, however, is ready to tackle the challenge with specific, tangible solutions.
With rapidly changing conditions that put more drivers and more complex cars on the road, it’s no surprise that auto claim volume and loss costs have increased substantially in recent years.
Find the Needles in the HaystackOne of the rapidly changing conditions is the abundance of new data from new sources. Data from sensors in our cars, consumers’ mobile devices, repair facilities diagnostic tools—we have no shortage of opportunity to look more closely at claims and repair details. But therein lies the problem. While this data revolution offers unprecedented insight, without the proper tools to quickly sort and find meaningful information in the context of a claim, we are left to manually search for needles in an ever-growing haystack. This is where A.I. is ready to help. Artificial Intelligence technology can find patterns amidst massive amounts of data that would otherwise escape our attention. With A.I.-enabled solutions, carriers can identify claims that need closer attention, like finding patterns in repair/replace decisions that produce better results, giving them the power to focus resources where they are most impactful.
Assist and ExpediteBetter still, once the meaningful data is identified, A.I. can help to elevate the right information in a way that assists and expedites workflow processes. By leveraging A.I. and visual computing to analyze photos, for example, A.I.-enabled workflow solutions can use machine learning technology to minimize estimate errors and maximize reviewer efficiency. One such initiative is the Mitchell Assisted Review project, which was launched in October 2016 to accomplish exactly this goal. By utilizing millions of damaged vehicle photos, computers are “trained” to recognize vehicle damage and use computer vision to double-check repair vs replace decisions. This will help carriers achieve better estimate consistency, maintain estimate quality and be more selective about sending appraisers into the field, all while improving cycle times and productivity.
Keep it Simple StupidAs with any new technology, however, A.I. adoption as a standard part of the claims workflow will only reach critical mass when implemented with usability and practicality. As part of the Mitchell Assisted Review project, for example, User Experience (UX) designers are working hand in hand with artificial intelligence technologists to design solutions that highlight repair/replace outliers in a way that makes sense within the claims workflow. Reviewers have neither the time nor the inclination to take on complex and time-consuming new technology tools, so for A.I.-enabled solutions to be effective, estimates in need of review must be easy to spot, easy to understand and, most importantly, easy to act upon within the review workflow. Making A.I. technology easy and practical for the claims professionals who will interact with it on a daily basis is key to unlocking the technology’s full potential. The more the solutions are used, the smarter the technology becomes, thus allowing for constant improvements in efficiency, accuracy and consistency while providing increasingly better insight with which to inform estimating guidelines that reinforce trust and acceptance.
By leveraging A.I. and visual computing to analyze photos, for example, A.I.- enabled workflow solutions can use machine learning technology to minimize estimate errors and maximize reviewer efficiency.