There has been a lot of talk lately about “machine learning,” and how it enables computer systems to evolve algorithms, without programmer intervention, as these systems take in updated knowledge and insights. In other words, algorithms are capable of learning and making appropriate shifts in the predictions they produce.
Is the insurance industry ripe for machine learning? Where will it produce the greatest benefits? Analytics are already being applied in data-intensive areas ranging from fraud detection to applicants’ risk profiles, so there already is an artificial intelligence aspect of today’s operations.
There’s plenty more where that came from, notes Steve Anderson, a well-regarded expect on all things insurance technology related.
Recently, Anderson explained, he developed a machine-learning algorithm that took on many of the aspects of an agent-to-potential-policyholder discussion:
“I spent the day at a small company that is developing an expert conversation engine tool that will allow anyone to create a ‘guided conversation.’ Using their platform, I was able to create an online conversation that created a guided conversation that answered the question, ‘I bought a new boat. Does my homeowner’s policy cover it?’ Building the response to this question took about an hour for me to complete. The questions I asked included: ‘What type of boat did you buy?’ ‘Does it have a motor?’ ‘If yes, what is the horsepower?’ ‘How much did you pay?’”
Anderson adds that he was able to build this algorithm in about 30 minutes, and was even able to provide information about additional insurance coverage.
Does this mean agents should worry about being replaced by algorithms? While these systems are getting easier and cheaper to build, insurance agents need not be concerned for their jobs, Anderson continues. As he explains, machine learning may enhance and enrich insurance engagements, but human agents will always be needed, because many insurance transactions involve much more than a simple exchange of data and money. However, for “for insurance agents who are simply ‘order takers,’ machine learning will likely be a threat.”
As I’ve mentioned in previous posts, insurers have the ultimate “social network” in their agency and broker networks. They have contact with policyholders that is more intimate than anything algorithms can capture. For example, a policyholder may have unique requirements that aren’t addressed within lines of code. However, with information technology and machine learning at their side, agents have a powerful set of tools that can enhance their relationships with policyholders.
Anderson spells out how enhanced systems can improve agents’ performance:
• Machine-learning systems can help “capture the knowledge, skills, and expertise from a generation of insurance staff before they retire in the next five to 10 years.”
• Machine-learning algorithms provide a way to reach the upcoming generation of consumers. “These consumers expect to be able to get answers anytime — not just when an agent’s office is open.”
• These enhanced systems help “provide consistent — and correct — answers to common insurance questions.”
• Having machine-learning systems at their disposal enables agents to elevate their roles in the insurance sales and policy administration process. Systems can handle the more routine day-to-day questions, and agents can invest more time and effort “engaging clients at a deeper level that requires more in-depth expertise.”
• Machine-learning algorithms provide deeper levels of knowledge and interaction on a 24x7 basis, versus a simple FAQ page on a website.
The bottom line: algorithms will eventually displace a great deal of the routine questions agents may need to answer on a day-to-day basis, and even handle application processes. But insurance will always be a very personal and relationship-driven business.
