Other parts of this series:
Artificial intelligence presents insurers with many new opportunities for efficiency and accelerated growth, but success depends upon determining both work complexity and data complexity for the tasks undertaken.
As I discussed in my previous blogs, artificial intelligence in the form of cognitive computing, robotic processing automation and other technologies presents insurers with many new opportunities for efficiency and for accelerated growth.
The question for insurers is whether to use AI to automate processes or to augment the workforce and make it more creative and effective. Many insurers are adapting AI to automate basic underwriting transactions and renewal processing. AI also provides consistent, low-cost performance in making underwriting eligibility decisions.
Insurers are exploring the use of AI to augment the expert workforce in areas including risk management, client and/or prospect discovery, coverage recommendations and fraud detection. These are clear examples of what we have termed work complexity, where individuals apply judgment and interpret information.
Use of AI is growing in areas characterized by data complexity. Here AI helps interpret large volumes of unstructured, rapidly changing data; for example, AI can help claims adjusters analyze the post-collision value of a vehicle to establish a true market value. Similarly, AI can help interpret photographic imagery from drones or manned aircraft to estimate property damage in the wake of a catastrophic event such as hurricane or tornado.
Data complexity and work complexity intersect in some areas such as product development and innovation, which place a heavy emphasis upon human judgment and experience. Even here, however, AI is being used to support innovation in areas such as home health analysis, customer personality profiling and “visual telematics”, such as combining the analysis of body movements of a driver with telematics data to establish levels of risk and to use such data for pricing and underwriting decisions.
In my fourth and final blog on AI, I will look at a use case involving various elements of AI in improving the speed and accuracy of underwriting for homeowners’ insurance.