AI, technology and the human element
AI detects patterns — people detect motives
With solutions like PromiseXP, Swiss Re is supporting clients with AI-driven automated tools for fraud detection. In the final section of our interview with Erick Sloeth, we discuss how important it is to have human experience, expertise and decision making at every stage of the journey.
Tim: How do you see the current situation investigators and technological solutions working side by side?
Erick: Experienced team members will be able to guide you with how to apply the technology. All of our clients have people dealing with fraud on a full-time basis. Including those people in any kind of tech development work is always a good idea.
A great example, we recently had a claim of a car accident where the individual was supposedly in an accident with a tanker truck and there was a big fire. We used the tech to conduct a reverse image search. Then our experts spotted that the photo evidence wasn't from 2025, but it was rather from 2014 and completely disassociated with the particular claim.
Tim: What about efficiency gains from new technologies and AI?
Erick: For me personally, I think it's fantastic that I'm spending more of my time being able to focus on the detection of fraud schemes and things of that nature. That's what I want to do. I don't necessarily want to spend my days typing out or doing something that could be kind of streamlined.
Take translations, for example. With the amount of international work I see, having the ability to quickly get an understanding of what something says allows me to apply my skill set appropriately.
We also need to be careful. We have AI, but so do fraudsters. There's a risk of gaming of the system to determine where the thresholds lie. This is where the human fraud specialist is really going to pay off. We're catching the stuff coming through now and setting ourselves up for protection in the future.
Tim: What are the human elements to consider when detecting fraud?
Erick: There may be messages that AI tools and technologies may not pick up on because they're more data-driven or more objective. So, you shouldn't lose out on the subjectivity and the human-to-human element that's critical for determining and detecting red flags.
We are advocates for continuing behavioral training. Measures like suspicion training that can help staff find red flags.
Tim: What is the mindset of people when Fraud is discovered?
Erick: I think it's showing up in areas of opportunistic fraud where there are probably certain demographics that are more easily able to rationalize the behavior within themselves. Exaggeration of data at the point of claim or perhaps withholding some information at the point of underwriting.
These rationalizations of things like, "it's only a small difference or the insurer can cover that". There's that difference in behavioral motivations as well for opportunistic fraud to look out for. There are strategies that we can build into underwriting and claims processes to encourage honesty and allow for ease of accuracy in disclosure too.