The Super Insurance Agent

Gone are the days of the door-to-door insurance salesman. For a time, his personable sales approach served customers and insurers well, but a new breed of consumers and an evolving market landscape have obsolesced traditional agent tactics. The insurance agent of yesteryear is fading away, and the field isn’t attracting much young talent. In short, the profession is in dire need of revitalization if it is to remain a central stream of revenue. Fortunately, technological advancements, rather than obsolete the agent, hold the promise of breathing new life into the field.

The agent profession has regressed into a low-skill data entry job, and the sales savvy within the agent force is going to waste. In light of this, one might think the obvious solution to the ‘agent problem’ is eliminating the outmoded insurance agent. But advances in consumer data gathering, coupled with machine learning technology and AI, can not only revitalize the agent profession but perhaps generate a whole new class of what I dub ‘super agents.’ The super agent is the result of an elegant combination of cutting-edge technology and good old-fashioned human persona. The super agent uses his natural born salesmanship in cooperation with AI and machine learning programs that provide him with a steady stream of relevant data on existing clients and market trends. The result is a super agent that upsells and cross-sells at a blistering pace—the holy grail of the insurance industry.

We know that most insurers are unhappy with the upselling and cross-selling of their agent force, but figures for upselling and cross-selling rates are hard to come by. One survey of 11 companies found that on average they cross-sell to just 23% of customers with the low-end being 1% and the top-end 37%. However, these statistics should be taken with an enormous grain of salt since the sample size was far too low and not random. In fact, the average cross-selling rate is probably significantly lower than the still-too-low one just cited. Anyways, one-off policy sales don’t build customer relationships, brand trust, or significant revenue. But insurers who can find novel ways of gathering customer data can convert those data into information that their agents can crucially leverage.

Allow me to paint you a rough picture of the super agent model. Imagine that super agent Sam sells a small term life insurance policy to Jones, who is young in his career and without dependents. The normal agent would be left to make a cold, somewhat arbitrary follow-up call a year later, if he even remembers to do so. But super agent Sam gets notified that Jones is now married, and his wife is expecting. He now has an excellent reason to call Jones up and recommend an increase in coverage or perhaps a conversion to permanent life insurance. Jones buys and is the happier for it, knowing his family’s future is more secure.

In this example, we see multiple points of agent enhancement: the agent’s sales abilities are enhanced by having personal information on the customer with which he can influence his decision and also by having the right timing for calling the prospective buyer; the customer’s experience is enhanced, for he feels that he is getting personalized service, perhaps even that his agent has his financial security in mind; and of course there’s the enhancement of the agent’s efficiency, for he can use the information the program feeds him to narrow his focus to only those most likely to buy a policy (in fact, the program could automatically queue up leads for the agent, where the persons queued would be those the program has algorithmically determined are most likely to buy).

Life event data of this kind can be sourced in different ways but few insurers are taking the steps to do so. Some personal data are publicly available via sources like the DMV’s Motor Vehicle Reports, whereas other data are readily purchasable from corporations such as LexisNexis and the Medical Information Bureau. Additionally, there’s an opportunity to collect ongoing data by providing consumer-facing web and mobile applications. These apps can go beyond policy management and provide many use cases for acquiring dynamic data on policyholders. Open authorization logins, rewards programs, and social features within such applications present opportunities to collect customer data that insurers have never had before.

In our example, Jones is not surprised that Sam knows of these life events, not merely because he volunteered access to his Facebook, etc., but because of the rapport Sam built with him upon their first meeting. Indeed, the millennial generation has shown a remarkable willingness to exchange once closely guarded personal information for the rights to products and services. Of course, personal information can be obtained more indirectly, and the best recipe for a super agent is probably a mixture of both volunteered and indirectly gained information, resulting in a stew of personal details for the super agent to use according to his sales expertise.

Clearly, the case delineated above does not work for all types of insurance, or all types of insurance agents, but this is all well and good, for the types of insurance it doesn’t work for, like car insurance, can be completely automated: policies can be bought through a mobile app or other internet platform, chatbots can field customer questions, and policy management can be done via the same platform on which the policy was bought. Thus, the super agent, or any agent for that matter, is irrelevant for certain types of insurance, yet indispensable, that is, indispensable if one wants to maximize his profits, for other types. Life insurance is a perfect type, so is home insurance, but the myriad specialty insurances will really be where the super agent’s bread is buttered. The customer who will be cross-sold and upsold the most is an upper middle class to wealthy individual who owns many insurable items. If an agent could cultivate enough customers approaching this ideal, and it wouldn’t take many, then his cross-selling and upselling percentage would be more than satisfactory. Thus, the super agent would be reaping more profits from a smaller subset of customers than ever before; he would, of course, still have the trickle-in customers accounting for most of his pecuniary gains, yet the small collection of customers to whom he regularly cross-sells and upsells would bring in profits in great disproportion to their head count. Similar models have been highly successful in other industries, for as long as a company cultivates a core base of customers it regularly sells to, that guaranteed income allows the company to maintain profitability even in the face of downward market trends.

Despite the foregoing illustration of the super agent model and its benefits, one may still not be convinced pure automation is inferior, asking, “Well, we know the current model isn’t working, given the lack of cross-selling and upselling, but how do we know the super agent model is more profitable than implementing pure automation?” While it is true that all insurance could be automated (excepting, perhaps, specialty items, which may be the only argument one needs against pure automation), it is more plausible than not that this would be a much less profitable avenue than the one the super agent strolls. For, the multiple points of enhancement listed above would simply not be present in a purely automated insurance industry as there would be limited customer satisfaction and thereby limited willingness to buy due to the whole impersonality of the affair, and there’d be no cultivation of, what we may call, ‘super customers.’ Furthermore, since the agent’s income can be primarily commission-based, the bulk of the pecuniary subtraction his employment incurs will be in direct proportion to his sales thereby resulting in a net gain for the insurer. Thus, given its superior profitability and customer satisfaction, it’s clear that the super agent model is the best option for optimizing insurance sales moving forward.