Although artificial intelligence (AI), IoT and big data have become inescapable buzzwords, the notion that software will perform functions typically associated with the human mind is fast becoming a reality.
In China, several examples of AI use cases spanning distribution, claims and loss adjusting have already emerged and China’s internet giants (Baidu, Alibaba and Tencent) are keen to enter the industry with both distribution and underwriting activities.
The case for AI deployments within insurance is also bolstered by the fact that insurers have petabytes of data. This, combined with government support, is fueling the deployment of AI in distribution, claims and loss adjusting.
1. AI in Distribution:
Although still in their infancy, chatbots and speech recognition features will soon represent a new distribution channel for insurers. The Digital Insurer’s Rick Huckstep has already examined the merits of speech recognition and bots here and several efforts are also underway in China including;
- Kitt.ai: Kitt.ai develops a conversational language plug in called ChatFlow that allows developers to add natural language processing functionality into apps. Recently acquired by Baidu, Kitt.ai built its reputation by perfecting trigger words that activate pre-determined actions in apps.
- iFlyTech: iFlytech is China’s most advanced speech recognition provider. Based in Anhui province, iFlytech’s voice assistant boasts a 97% accuracy rate for Mandarin recognition, significantly out performing that of Siri and other Chinese offerings.
- Innolife is a text based chatbot built for WeChat. It’s AI component is based on a decision tree of user responses that is constantly added to as users engage with the bot. This allows Innolife to tailor health, life and disability insurance policies to users needs.
In addition to this active startup scene, several established insurers are also imbuing speech recognition and bots into their customer service and distribution efforts. Again, Ping An is the most advanced with a dedicated AI team working towards reducing reliance on its call centers. Ping An’s goal is handle text based policy queries on WeChat instead of through customer service agents. The bot was developed by first collating all policy holder questions into a central database, which is then continuously iterated upon in real time for Ping An’s 80m policy holders.
Elsewhere, Taikang is taking a very literal interpretation of chatbots with its TKer robot, a life sized chatbot designed to replicate the role of agents in tier one cities.
Through scanning the user’s national ID card, customers can query policy features, apply for renewal, and engage in basic financial planning that includes ancillary products from Taikang. The TKrobot is also equipped with speech and facial recognition functionality for repeat users.
Finally, as TK robot’s speech recognition capabilities are still developing, the TKer has also been equipped with a VoIP connection to Taikang’s customer service departments.
Ultimately, the use case for interactive financial planning terminals in banks, real estate outlets and shopping malls is compelling. Just as ATMs have replicated the functions of bank tellers, Taikang’s TK robot and the broader advancements in speech recognition is the start of a paradigm shift in the distribution of life/health insurance.
2. AI in claims and loss adjusting:
The opportunity for AI to help validate claims and perform basic loss adjusting services is obvious. However, before detailing the key efforts currently underway, its important to note that the level of historical claim data required to train an AI based solution is difficult for startups to acquire, a challenge that Rick Huckstep has already examined here.
Despite this, several players are endeavouring to produce AI that can improve the validation and settlement of insurance claims in China. Of these, Beijing based 4Paradigm is the most notable for its pioneering work with PICC in tackling the claims fraud that continues to plague China’s auto insurance sector in particular. PICC, a state owned carrier and China’s largest property insurer, has no shortage of claim data from which 4Paradigm can train its neural network (the software that identifies suspect claims). This has given 4Paradigm a head start in training its AI layer to recognise anomalies and outlier data points in claim applications.
As PICC and other state owned insurers are busy partnering with InsurTechs to capitalise on AI, China’s pre-eminent digital insurer, Ping An, has chosen a more independent path. Ping An’s recently announced ‘Smart Fast Claim’ has already processed over 4.9 million auto claims in the first half of 2017 and intercepted almost 3 billion rmb of leakage (the gap between optimal and actual claim settlements). Additionally, the ‘Smart Fast Claim‘ has reduced the time taken to settle claims processed from three days to 30 minutes.
Another vital element of claims processing is loss adjusting, a function that impacts insurers profitability directly. In this case, China’s internet giants (Baidu, Alibaba and Tencent) are contributing with AI based loss adjusting platforms. Of these, Alibaba’s is the most sophisticated in terms of AI depth and takes seconds to assess the exterior damage of auto claims. Yu Zhu, general manager of China Taiping’s Property Insurance division has said of the system: “Ant’s team achieved in one year what we have been working on for the past five”.
Although this is a radically different method of assessing claims, it’s worth noting that even the most sophisticated loss adjusting solutions will still have limitations – internal damage that is invisible to image recognition for example, or prior generation models that look similar but have structural changes.
Furthermore, this loss adjustment platform from Ant is currently being offered as a free white label solution to China’s auto insurers. Although making hard earned innovation publicly available to competing insurers may come as as surprise given the intensity of competition from in China, the fact that billions of images are required to validate claims and support repair estimates means that making the platform publicly available will extend the lead that Ant has forged in this space. Ultimately, Alibaba’s contribution here is a great example of AI based solutions that reach deeper into the value chain than purely distribution focused efforts.
Finally, the often over looked operations department also boasts some interesting AI applications from China. Here, customer authentication, CRM and security services are all feeling the effects. Again, Ping An is breaking new ground with its ‘biometric authentication’ feature that offers policy holders the option to enhance their account access through face and voice recognition. The feature, delivered through Ping An’s ecosystem of apps, scans the structure of the face (especially the nose and eyes) with a higher degree of accuracy than the human eye and is also being used to verify the identity of insurance applicants or agents.
Additionally, the biometric authentication feature also helps Ping An to form a relationship with pending policy holders (those who have applied for life/health insurance but are waiting for the policy to be activated). Since launch, the rate of customers who withdraw their application during this window has dropped to 1.4% against an industry average of 4%.
AI based technologies hold exciting and far-reaching potential to improve healthcare, the environment, and education. Although Ping An currently has the upper hand in terms of access to data upon which AI solutions can be trained, the examples above demonstrate the wide ranging efforts currently underway in China, and, with government support and unmatched scale, China has and opportunity to lead the development of AI based technologies globally.
To read the latest insurance innovation news and features in Chinese on WeChat visit our page at: