Leveraging AI: A road map for claims executives
There is some hesitancy regarding the adoption of AI technology and uncertainty as to what approach to undertake.
Whenever I speak with claims executives, they tell me how they are being inundated with information about artificial intelligence (AI). Many are reluctant to adopt this technology although they know delaying the investment can result in diminished business in the future.
According to “Making AI Responsible — and Effective,” a 2018 Cognizant AI survey, “only 51% of insurance executives said that AI technologies were extremely or very important to their company’s success today, which is lower than for any other industry.” Why does our industry lag behind others? Much of the hesitancy regarding the adoption of AI technology is a fear of the unknown and uncertainty as to what approach to undertake.
AI implementation
Beginning an undertaking as process altering as the use of AI technology can be overwhelming. As a starting point, you have to set the strategy for the use of AI technology in the claim operation throughout the organization. This means you must become familiar with AI technology and terminology by reading articles and white papers and attending InsurTech conferences.
It’s also helpful to find a trusted advisor, either in-house or from the outside, who can explain AI technology in terms that non-technology people can understand. Only by understanding the technology can you make a realistic appraisal of the business needs and develop a strategy to meet those needs. This means finding areas where the AI technology can be applied, understanding the risks and rewards, and having the resources needed to do so.
Once a strategy is formulated, create an AI team to recommend possible solutions. The AI team charter will reflect your strategy for using AI technology in claims and research the possible solutions on the market as well as the vendors who supply them. Integral to this consideration is the creation of a pilot program that is of sufficient scale so as to be meaningful. The AI team should also consider how the technology will be integrated into the claims process and determine how success for the integration and pilot program will be measured.
To illustrate, assume you decide that remote damage assessment to property in a catastrophe is an area to address with AI. The team would analyze the impact this AI solution would have on various internal and external stakeholders. Did the technology improve the customer experience? Did the AI technology show a marked improvement in adjuster response time and the time it takes to close a claim? Did the technology improve the accuracy of reserves for reinsurance purposes?
Begin the analysis with the external stakeholders — the policyholders and agents/brokers. These are the people who will be most impacted by the implementation of new technology. Many will fear its use so you should ask if the policyholder and agent felt they were getting personal service even though the AI technology was used to assess the damage. Develop a communication strategy to help with the implementation of the AI technology that addresses policyholder and agent concerns
Choosing a solution
Once the concerns of the stakeholders are adequately addressed, the AI team can move on to the task of researching and selecting a technology solution. Most insurers have an established protocol for the selection of a vendor and once selected, the AI team can focus on a pilot program to “road test” the solution in a limited way under real circumstances. Does the solution perform as promised? How easy or difficult is it for staff to use? Does it truly deliver a high-quality policyholder experience? Performing a random audit of claim files can help answer these questions.
A successful pilot program will also highlight the areas internally where it will be necessary to build skills and organizational capabilities. It will be necessary to review the claim handling process to find elements no longer needed such as establishing a reserve for outside adjuster expense as well as new elements that now have to be included such as the roof style, roof dimensions and type of shingles. Existing business processes will have to be optimized to take full advantage of the new AI technology solution. This means creating a talent and technology infrastructure that can utilize the AI solution to its fullest potential.
At the end of the pilot program, evaluate the success of the AI solution after every use. The AI team should create success criteria for the policyholder experience, agent/broker experience, business use, and other stakeholders. The team should also design criteria that measures return on investment. For example, is the claim turnaround time noticeably shorter than before? If so, by how much? Are policyholders happy with the time frame? Has earlier intervention mitigated the extent of the loss? Has claim reserve accuracy improved? Has there been a reduction in operating costs? Answers to these questions will determine how successful the investment is.
Developing and realizing a strategy for applying AI to claims business challenges may seem like a daunting task, but it need not be. In many ways, developing the strategy and selecting a vendor is the easy part of the process. The next step, implementation of the pilot program, will fall to a claim manager who will have the responsibility to change the internal process to make use of the technology. This phase in the process will be the focus of my next column. The claims manager will have to report the results of the pilot program and hear the comments from policyholders, agents and staff. How this burden is handled will make or break the return on investment.
Donna J. Popow, Esq., CPCU, (donna.popow@verizon.net) spend more than 30 years in claims before becoming a consultant to the property/casualty insurance industry. She has managed catastrophe operations and written extensively on claim-handling procedures.