Fundamentally, paying claims is what insurers do; indeed, it constitutes the lion’s share of their outgoings. For P&C insurers, for instance, that typically amounts to 60-80 percent of costs.
The simplicity of this premise masks a great deal of complexity, of course, and insurers need to balance three elements that are often in opposition to each other:
- Contain payments losses – pay what’s appropriate, and only what’s appropriate
- Maintain customer satisfaction – customers generally don’t have much contact with their insurers except in claims situation, which is the “moment of truth”
- Hold down the cost of claims administration
While it’s perhaps easy enough to balance two of these, any combination often comes at the expense of the third. So, you can keep customers happy and administrative costs near zero by paying every claim in full; however, your loss ratio, the key KPI, will go through the roof. Alternatively, you can manually process each claim, establishing with absolute certainty you are paying only those you should: your payment losses KPI will be superb, but you’ll have high costs and unhappy customers.
Getting this balance right has been the industry’s challenge since its inception, with two out of three the best it could do – until now. Technology is remaking the claims pay-out space.
First, pay up
At Accenture, we’ve long sought to help clients industrialise the claims management process, making it operate like a factory with only as much time spent on claims as is necessary to pay what’s right. That means automating the claims processes, then branching out only if needed. The constant goal is to optimise the balance between time spent and the impact on the pay-out outcome.
AI and analytics have revolutionised what’s possible. As I wrote before, in China, the scale of the market means insurers were forced to pursue a digital route. As a result, they have become global leaders in the use of data, artificial intelligence (AI) and analytics – streamlining the full range of insurance processes, from underwriting to paying claims. Today some Chinese insurers have inverted the claims process: their default is to pay, allowing them to balance all three aspects. Here’s how they do it.
The key is to build a straight-through-payment process as the default, for which technology has provided progressively better solutions. However, this requires a cultural shift, with insurers changing their mindset from finding any reason not to pay, and instead of paying every claim quickly except for those where there is a sound reason to delay or stop payment.
Step one is to implement more sophisticated workflow solutions so that, instead of escalating all claims above, say, US$10,000 to a supervisor, they check only those claims that deviate from a predefined approach, i.e. those that raise data-analytics flags. Indeed, our studies show that “leakage” (the costs incurred administrating or paying out claims) is proportionally higher with small, high-volume claims that are “uninteresting” than it is with the large ones that typically are scrutinised.
The second is to use analytics to compare each claim’s data against its peers, looking for outliers. Why does this windscreen replacement cost three times the average? Why is this insured person making a third claim between the same two people in a year? There might be good reasons, but there might not be. This is better than fixed rules systems, as they are too generic. (One client, for example, saw 80 percent of claims red-flagged, with operators consequently clicking away every flag as they hadn’t time to see whether they were valid.)
The third is to use AI at key decision-points – as Ping An does with damage recognition where the insured sends photographs of their car after an accident, and the system estimates the likely cost. This approach is also helpful for more complex areas like hospital claims. Critical illness insurer Xiang Hu Bao, for example, has fully automated its claims adjudication system leveraging AI and blockchain to enable digital evidence submission.
AI and predictive analytics can be used at other decision-points to enable automated standard-path payment or to stop the process. These points include coverage match, liability assessment, fraud detection and final payment decision.
Lessons for all
Insurers elsewhere can learn from the approach pioneered by Chinese insurers to increase claims accuracy, reduce leakage and boost customer satisfaction.
Data is crucial. Many insurers try to integrate multiple systems, paper reports and information held on external databases. That’s close to impossible at scale without the technological tools that pull data from different sources and place it in structured databases – for example, using AI and optical character recognition (OCR) to extract data from written documents and feed it into structured databases, or to analyse legal documents or police reports of accidents.
However, once that data is in place, insurers can apply AI and analytics to drive automation, making pay-outs the default, and ensuring that claims adjusters spend their time more valuably and have more interesting work.
Insurers can also use technology to boost the second element – customer satisfaction.
Source: Francois Metzler, Accenture
Share on linkedin
LinkedIn
Share on twitter
Twitter
Share on facebook
Facebook
