Adoption of AI for Auto Claims Adjudication

By January 27, 2020 January 28th, 2020 AI, Blog
AI for Auto Claims

Insurance claims adjudication has come a long way since analytics was first used to identify authentic and straight-through claims a few years ago. Nevertheless, the auto insurance industry faces many challenges in adjudicating claims today as we enter the new decade; key among them is the rapidly rising Loss Adjustment Expenses (LAE), given the trend of increased severity and frequency of the incidents.

There is an increased urgency among industry experts to find a solution in order to control expenses and improve Net Promoter Score (NPS). The industry is increasingly looking at AI as a possible technology enabler; as AI has already been making a significant strides in underwriting, risk scoring and marketing. While claims processing has been relatively slow in AI adoption, all indications show that this will change significantly in the coming years.

Unique Challenges
The auto claims adjudication process faces many challenges like delayed reporting, longer cycle times, human error in assessment or filing, fraudulent claims, customer dissatisfaction, and a lack of transparency in the process from the customers’ viewpoint.

Each step in the lifecycle of a claim after FNOL (First Notice of Loss) can take days or weeks to process. The effort required by adjusters in triaging and investigating various aspects of a claim requires significant data crunching and analysis. This is especially true for:

  1. Checking the authenticity of a claim
  2. Accurately assessing the damages

Both of the above factors are key to adjudicating the claim accurately, efficiently and quickly, with a direct impact on expenses and Net Promoter Score (NPS).

Let’s look at how an AI system can enable adjusters and the overall adjudication process to be more effective:

Checking the authenticity of the claims:
Right at FNOL, the AI-powered solution can compare the claims data points with historical data and run it against known fraud markers. The solution can assess the damage, check the incident report, verify data against the policy terms, and run fraud detection algorithms to identify the nature of the claim accurately. The claim is then assigned a score and its propensity of being genuine or fraudulent is determined. The adjusters can then simply verify/validate the findings of the score without extensive triaging and take appropriate actions. Complex or potentially fraudulent claims can be routed more quickly to the experienced adjusters or SIU investigators, depending on the set parameters. Whereas, simple claims can be queued, up to be auto-adjudicated after verification.

Accurate assessment of damages:
One of the most time-consuming activities in the whole lifecycle of a claim is the accurate estimation of damages, this not only impacts the final payout, but critically important for an accurate reserving amount.

AI models, trained on images of wrecked cars and incident reports, can quickly produce a fairly accurate draft estimate, which can then be used for reserving and validated against an estimate from a body shop to arrive at an accurate and verified final estimate.

As real-time image recognition becomes more effective and less expensive, it may soon be possible to offer even more sophisticated services like on-the-spot image capture, auto-trigger of towing and repair services, a transparent incident report including all relevant data, and an estimate of the settlement amount right at FNOL stage.

Enhancing customer experience and improving customer satisfaction scores:
Surveys show that while one in five consumers do not like to answer too many questions at First Notice of Loss (FNOL) and prefer the self-service claim options. But lack of knowledge of insurance policy terms and inadequate coverage add to the delays and fuel even more dissatisfaction.

AI models can play a significant role in both highlighting relevant insights about a claim quickly, but also generate customer alerts based on those insights. With an accurate, transparent and timely reporting process driven by AI, the customer feels more confident about the fairness of the claims process and are more likely to accept the settlement offer. The settlement amount would be in line with the right parameters, and hence customers’ interests are safeguarded, and human errors and biases are eliminated.

The Future of Claims Settlement: Next Level of Automation
Going a step further from automated claims report filing and incident reporting, another advantage that AI can deliver is automated claims support. While new customers or complex cases may require a human touch, straight-through and other less serious cases may benefit from AI-based chatbots that can coordinate the entire process and keep the customer updated about the status of their claim.

Such an automated system of claims support is not heavy on resources and is valuable for delivering much superior customer experience. While 100% touchless claims may not be possible today, AI-powered chatbots can definitely free up resources and reduce human errors to a high degree.

To sum up, there is tremendous potential to deliver a seamless insurance claims, customer experience through AI technologies; it not only helps improve Net Promoter Score (NPS), but also helps insurers reduce their Loss Adjustment Expenses (LAE) and improve cycle times, creating a win-win for all parties involved.

AI for Auto Claims

Maneesh Madan

Maneesh is a CEO of InsurAnalytics.ai – Delivering AI powered insights to P&C Insurers. He is responsible for the organization's vision, strategic direction and alignment of organizational resources to enable a customer centric approach. Read More Posts

Leave a Reply