In the past few decades, insurance companies have collected vast amounts of data relevant to their business processes, customers, claims, and so on. This data can be unstructured in the form of PDFs, text documents, images, and videos, or structured, organized and curated for big data analytics.
Mckinsey claims the state of the artificial Intelligence market in the Insurance industry will be impacted by four factors: an explosion of data from connected devices, open source data ecosystems, advances in cognitive computing technologies, and an increased prevalence of physical robotics. We previously covered how the top 4 American insurance firms are using AI, and in this report, we specifically cover AI-based business intelligence (BI) solutions for the insurance industry. As of now, numerous companies claim to assist insurance professionals in aspects of their roles from developing personalized insurance products to claims processing.
We researched the space to better understand where AI-based business intelligence applications come into play in the insurance industry and to answer the following questions:
- What types of AI-based business intelligence applications are currently in use in insurance?
- What tangible results have AI-based business intelligence applications driven in insurance?
- Are there any common trends among these innovation efforts? How could these trends affect the future of insurance?
This report covers vendors offering software across three applications:
- Data Management
- Insurance Product Personalization
- Claims Modeling
InetSoft is a New Jersey-based company with 65 employees. The company offers a suite of software which it claims can help Insurance companies make better use of their enterprise data, such as that from SAP and CRMs, using machine learning.
InetSoft states that insurance companies would require a team of data scientists and domain experts in insurance to integrate their software. Insurance firms can also hire an Inetsoft consultant to help with organizing their business data (such as historical claims records) in a format that is readable by the software. Then, the software analyzes the data using machine learning algorithms and allows the segmentation of this data.
The software automatically clusters and predicts forecasts for the data to help insurance companies acquire insight they can use to make decisions. For instance, the software might classify a group of claim applications based on the geographical locations of the applicants. The system might then provide visual representations of the analyses in the form of graphs and charts on an interactive dashboard.
Below is a short 2-minute video demonstrating how the company’s software works:
InetSoft seems to have several case studies in which it worked with insurance firms in deploying business intelligence software, although it should be noted that this software was not AI. We could find no case studies involving the company’s BI software.
InetSoft lists ING North America Insurance Group, Canal Insurance Company, and the Great American Insurance Group as some of its past clients. The company also claims to have worked in projects with NASA, the CIA, WHO, IBM AT&T, GE among others.
There does not seem to be any C-level executives on the team with an AI background, which is a red flag with regards to the legitimacy of the company’s AI technology.
GrayMatter Software Services
GrayMatter is a Bangalore, India-based company that offers a business intelligence software suite called Insurance Analytics (IA+), which it claims can help insurance companies centralize their data by creating a data warehouse for them.
Graymatter claims clients can input data from multiple sources into IA+, such as that from social media, CRM data, information on lead management, and recruitment data. Users can also upload insurance industry data into IA+ from GrayMatter’s sources, such as Life Asia, Policy Asia, and eBao.
Then, IA+ categorizes the information into relevant segments as determined by a team consisting of employees from both the client company and GrayMatter. These segments might include claims, actuarial, finance, and underwriting. The system then provides metrics on a dashboard that can be accessed through desktop of mobile platforms by the client’s employees.
Below is a short 4-minute video where Chandana Jayasooriya, CTO of AIA Insurance Lanka PLC, explains how they use the IA+ software:
GrayMatter claims to have helped Sircon Corporation, an insurance software provider, create a software tool that would allow the insurance firm’s customers to create their own business intelligence reports and data visualizations. Data scientists from GrayMatter worked alongside developers at Sircon to customize GraMatter’s Pentaho BI tool. According to GrayMatter, Sircon was able to provide analytics services for their insurance clients, including additional capabilities, such as automatic report generation, in Adobe PDF or Microsoft Excel. We could find no other mention of measurable results for Sircon as a direct result of the integration.
GrayMatter also lists Momentum Insurance South Africa and AIA Bhd.Malaysia as some of its past clients.
Although GrayMatter seems to employ a few data scientists, we were unable to find any C-level executives with an educational background in AI on the company’s team. Vikas Gupta, GrayMatter’s CEO, previously served as co-founder and CTO at Manthan Systems.
Insurance Product Personalization
Progressive is an Ohio-based insurance company with over 23,000 employees. Progressive is reportedly using machine learning algorithms for a predictive analytics application in auto insurance. Progressive claims clients who are part of the Snapshot program can get personalized insurance rates based on their driving history. The insurance firm claims the Snapshot mobile app has collected 14 billion miles of driving data. Progressive incentivizes Snapshot for “most drivers” by offering an auto insurance discount averaging $130 after six months of use.
Progressive claims users can install the Snapshot app on their smartphone. The company collects driving data for customers who sign into the app and leave their phone in the car as they drive. The app uses this data to improve the performance of the machine learning algorithm behind the software. Then, Snapshot uses the data to determine which customers are less likely to get into accidents.
Progressive then offers lower insurance rates to these drivers. Customers can save more on their insurance by following the company’s suggestions, such as limiting the usage of hand brakes, avoiding late night diving, and staying off the phone while driving. The system then provides customers with a personalized insurance rate and a dashboard to see their driving data.
Progressive exclusively offers Snapshot to its customers, and so there are no available case studies for the software.
Pawan Divakarla is Data and Analytics Business Leader at Progressive. He holds an MS in Civil Engineering from Georgia Tech. Diwakarla has served with Progressive in various capacities for over 14 years.
DataRobot is a Boston-based company with over 400 employees. The company offers Automated Machine Learning platform, which it claims can help insurance businesses Optimize business processes using machine learning.
DataRobot claims users can use their automated machine learning platform for applications like claims development modeling, life insurance underwriting, and direct marketing. Users can input social media, CRM, and claims data into the software through a dashboard interface.
Then, the software analyzes the data to predict future data points using machine learning algorithms which can identify patterns in historical data. For instance, the value of the claim might change drastically between the initial filing and full payment. DataRobot’s system can help insurers predict the final value of these claims to help clients make faster business decisions. The system then provides a graphical representation of the analysis on their dashboard.
Below is a short 3-minute video demonstrating how the DataRobot Predictive Modeling software at LendingTree works:
Although DataRobot seems to have case-studies for their automated machine learning solution in finance applications, we could find no evidence of robust case studies on the company’s website specific to the insurance industry.
DataRobot’s lists Evariant and DonorBureau some of their past clients.
Gourab De is VP of Data Science at DataRobot . He holds a PhD in Biostatistics from Harvard University. Previously, De served as a data scientist at Ginger.io.
Takeaways for Business Leaders in Insurance
The insurance industry checks off several boxes that might indicate readiness with regards to AI adoption, such as a large trove of historical data and the existence of large businesses with the capital and resources to successfully implement AI.
Based on the companies we researched for this report, we believe that larger insurance firms, such as Progressive, might have the resources to build and deploy their own AI solutions to improve business intelligence. These firms are investing in hiring and training multi-disciplinary data scientists, software developers, and insurance analysts.
For instance, Progressive is using AI to provide behavioral policy pricing to its customers. With Internet of Things (IoT) sensors providing health, driving, and lifestyle data, AI software can now personalize pricing for customers based on their behavior. What this means in real-world insurance business terms is that safer drivers might pay less for auto insurance (known as usage-based insurance), and similarly, people with healthier lifestyles might pay less for health insurance.
DataRobot has managed to raise over $124 million in funding and is backed by venture firm New Enterprise Associates. DataRobot might have managed to simplify their predictive analytics solutions to a point where users not familiar might be able to gain useful insights from the software. Users simply input historical data (such as claims records) through a web interface and can choose the target variable for which they want to forecast the data. The software lets users pick from a variety of built-in predictive models and deploy the one that is accurate.
Currently, data management for insurance firms using AI for insurance applications also seem to have traction. This might be down to the fact that although the industry has collected a lot of data, much of it is still unusable in their current format (such as paper documents or image files). Insurance firms also need to manage all their incoming documents, such as claims or application forms to ensure that the right documents are being delivered to the correct teams internal to the organization.
Report generation, summarization, and abstraction might be a low-hanging fruit application for business intelligence AI applications in insurance. We expect that certain standardized documents, such as leases or contracts, might be digitized more by insurance firms.
Business leaders at large insurance firms might need to be aware of the implication of having data spread out in physical locations across the world. This data might be affected by local regulations or even compatibility issues with data storage systems at each location. Even when all the data becomes accessible, a round of cleaning and parsing the data is necessary, and this whole process might take two to three months.
Businesses would also find that in some cases, regulations might mandate the signing of data sharing agreements between the involved parties, or data might need to be moved to locations where it can be analyzed. Since the data is highly voluminous, moving the data accurately can prove to be a challenge by itself.
Header Image Credit: Tobias and Tobias