A Tale of Two Sources

/in /by

It occurred to us the other day that one of the terms we don’t seem to hear much anymore is big data. Maybe that shouldn’t be surprising since the term refers to structured, unstructured, and semi-structured data that continues to grow in ever-expanding volumes at ever-increasing rates. So, we conducted an experiment with AI.

We put the same question to ChatGPT and to Perplexity. The question was this: What is the most common challenge for administering property/casualty insurance? We got two different responses, which seems to prove the point that even AI can’t keep up with the sheer volume of data to aggregate and manage it consistently.

Here’s what happened:

Behind Door #1

According to ChatGPT there were seven common challenges. They were:

  1. Underwriting Accuracy & Risk Assessment
  2. Fraud Detection & Prevention
  3. Regulatory Compliance
  4. Claims Management & Processing Efficiency
  5. Pricing & Profitability in a Competitive Market
  6. Technology & Cybersecurity Risks
  7. Customer Expectations & Retention.

Okay. That’s more than we asked for. But they all seem reasonable to varying degrees, depending on the software you use to administer the property/casualty insurance you’re writing.

Behind Door #2

According to Perplexity, there were five common challenges:

  1. Natural Disasters and Climate Change
  2. Economic Pressures
  3. Market Hardening
  4. Regulatory Challenges
  5. Operational Inefficiencies.

Those, too, seem reasonable to varying to degrees, even though they seem to view the industry from a higher, less technical level. But we were taken aback by this response to our Perplexity query:

The most common challenge in administering property and casualty (P&C) insurance is managing the increasing frequency and severity of natural disasters.

Yikes! We think the world of Aspire, our core processing software. And we’re proud of the development efforts that went into it and the refinements, enhancements, and improvements it continues to undergo. But being able to control the increasing frequency and severity of natural disasters is above our pay grade.

So, Then, What?

Since we haven’t yet equipped our software to manage the increasing frequency and severity of things like hurricanes, wildfires, and floods (we might need help from a higher power for that), a more strategic approach to writing property/casualty risks is in order. And while we can’t control the weather, we can work to create controls to better manage risk. Things like risk-based pricing, technological upgrades, and risk-mitigation partnerships with policyholders are increasingly necessary, along with improved tools for claims handling when losses do occur. But they’re not cure-alls. Until we come up with that cure-all, become clairvoyant, determine ways to be disaster-proof, or get AI to give us one consistent answer, we’ll just have to work together to do the best we can.

That’s pretty much the way we operate anyway.