Insurance agents must now embrace the use of artificial intelligence, but they won’t be replaced by it. And the use of parametric policies will soon increase sharply across property insurance lines because that coverage lends itself perfectly to data-driven AI programs.

Those were some of the major points outlined by Thomas Wetzel, an AI and cybersecurity consultant and educator for the insurance industry, at the Florida Association of Insurance Agents’ annual convention in Orlando.

“Insurance companies are going to become more comfortable insuring more types of things than they ever have,” said Wetzel, who previously worked for the Insurance Information Institute and now leads Thomas H. Wetzel and Associates, based in the Chicago area.

Wetzel

Parametric insurance, which can pay a set amount when certain parameters are reached, such as windspeed or flood levels, will soon be one of the fastest-growing areas that carriers embrace, thanks to the vast reach of artificial intelligence. AI can quickly analyze so much more data to accurately determine risk on a property or business, he said. Insureds will come to like parametric because payouts can be much faster, which can allow property owners to make repairs sooner and prevent further damage.

“The average car claim takes about six minutes. AI can do it in six seconds,” Wetzel said. “So, we’re going to be seeing a lot more parametric insurance, and it’s all because of AI.”

But human agents need not fear that they will be supplanted by “thinking” software, he noted. Clients will always want the human touch, a face-to-face meeting or a personal phone call from their agent.

Edmonds

Some areas, such as auto insurance on new cars, may bypass agents more and more in coming years as dealers use AI to automatically include policies at the point of sale. But for the most part, “there’s a zero percent chance of you losing your job because of AI,” he noted. “But it will change how you do your job.”

If agents rely too much on AI programs, such as ChatGPT or Microsoft Copilot, to do their jobs, then, yes, they will be marginalized to some degree, he added. But those that use it to extend their reach and expand the scope of insurance products they can offer will find it to be helpful and lucrative.

Here are some tips:

  • Use AI as a virtual assistant. This can mean using the software to draft emails, schedule meetings, summarize reports, or to research clients, risks and available carriers and coverage. Crafting social media postings also is a good use of the software.
  • Bring policyholders along for the AI ride. Be transparent with them on how it is being used.
  • One area AI is not good at, at least not yet, is binding, said Logan Edmonds, CEO of Tabernacle Tech, a cyber risk consultant who also was at the FAIA session last week.

“Binding is too complicated for AI at this point. No one has cracked the nut,” Edmonds said. “There are too many variables” and too much money at stake.

AI also has a little brother, known as cybercrime. And cybercriminals are using artificial intelligence to constantly improve their attack capabilities and exploit vulnerabilities, Wetzel said. To help combat cybercrime, insurance agents should train staff on the proper use of AI programs. Establish guardrails on what information can be put into the systems and what must be left out, such as personal identification info.

Wetzel and Edmonds recommended that agents work with a cybersecurity consultant to assess their vulnerabilities and develop safeguards to prevent getting hacked or to minimize damage. Multiple steps need to be taken, and old-school computer scans are no longer enough.

“The firewall on your computer blocks only about 20 percent of the threats,” Wetzel said.

Cybersecurity is a service, requiring constant updates, not a once-a-year product, the consultants said. Those services, including in-house information technology departments, must immediately make patches to fix software vulnerabilities without waiting.

Two-factor authentication is not perfect and is not always implemented correctly by businesses. Wetzel told of one insured business that used two-factor log-ins religiously – but only for certain transactions. The company’s insurer denied a cyber attack claim because the safeguard was not fully utilized.

New federal and state regulations on AI and cybersecurity are coming soon. “But the the best regulation is self-regulation,” Wetzel added. “Understand AI!”