Compliance AI – an Oxymoron?

Compliance AI – an Oxymoron?

May 24, 2024
Author: Joanna Hagelberger, Vice President Customer Success

According to a recent article published in the Claims Journal (https://www.claimsjournal.com/news/national/2024/04/10/322856.htm), “…35% of consumers polled supported using Artificial Intelligence (AI) to further fraud detection efforts.” Those percentages dropped steadily as the implications of those efforts became more personal.

Helping: People want good service. If the insurance industry can help me have a better experience by implementing AI, then everyone is all for it! Fraud makes insurance more expensive for all of us and I am not a bad person, so yes! Use AI to catch the bad guys! I don’t want to get ads encouraging me to insure the yacht I don’t own, so yes! Use AI to ensure the advertisements I get make sense for me. And I really don’t want to sit on hold for 20 minutes because I can’t remember if my auto policy covers rental cars. If a chatbot can answer that question quickly, then great – happy to chat.

Decision Making: Once the scales tip from making my world a better place to actually passing judgment, the article indicates the vote swung the other way. People are opposed to AI deciding the outcome of my claim or whether or not I should be insured (underwriting). These decisions could be influenced, they could result in negative consequences for the consumer, they couldn’t possible be made by a machine, could they?

Compliance: So where does compliance fall on this continuum? For better or for worse, consumers don’t care, but the industry certainly should. Can AI determine if a producer is authorized to sell a product? I believe the answer is sometimes which is why I think a subset of AI – Automation – is key to success in Compliance. Compliance is all about rules:

  • If you fill the application out correctly and meet X criteria, you can be automatically contracted.
  • Once you are contracted, we can automatically submit a background request using the required information in the application.
  • If you PASS the background investigation, you can automatically be appointed.
  • You will only be appointed in your resident state and Pennsylvania on day one (if you have a license in Pennsylvania). Any other Appointments will be held until you have business in hand (or whatever appointment rules you follow).
  • We will automatically send you a welcome letter advising you where you are now licensed/appointed to sell.
  • If your agents are captive, we can send out a background question form and automatically feed the answers into the system so all license renewals can be automatically submitted without ever being touched.
  • We can automatically remind producers when their licenses are due for renewal and/or how much CE they have completed.

This automation is amazing – it saves time and money, it elevates the work of your licensing team. But what happens when the “ifs” are not true? What does AI do with a “review” background investigation? An incorrect SSN/NPN? A yes answer on a background form? Automation is great when all the gates are open, but as soon as a gate is closed, human discernment/intervention is needed to ensure a compliant outcome.

Other Forms of AI: While most compliance professional won’t need to worry about these technologies in their daily work life, I thought a quick glossary might be helpful:

Predictive Analytics: Technology that finds and suggests similar products and experiences to the one you have just viewed (“I see you’re licensed in Ohio, wouldn’t you like to get a license in Pennsylvania?” I hope that never happens!).

Chatbots: A chat window that responds to key words entered by the user to provide intuitive responses. Intended to shorten wait times for common questions. If the chatbot gets confused, it triggers a human to respond.

Internet of Things (IoT): Physical objects that are embedded with sensors that are connected to the internet (like smart watches and monitors on cars).

Blockchain: A secure record of transactions that are shared across a network. While this isn’t really used broadly yet, imagine a world where every insurance company/agency could share the ability to do compliance checks. Unfortunately, data is not adequately standardized or publicly available (not to mention the astronomical security concerns), so we are long way off from benefiting from this technology.

Bottom line? When you hear all of the talk about AI in regulatory compliance, jump in with both feet and suggest all of the ways you can automate your processes to decrease manual errors, improve efficiency, increase time to revenue, improve staff retention (by decreasing redundant data entry!), and decrease costs. There’s nothing to fear in this brave new world.

480 316 Rhoads Online