The future of claims: Humans vs AI

James Sterling, head of claims, Kayzen Specialty.

As a cautious claims professional and former lawyer, deeply passionate about human-centric client service, my unconscious bias puts me naturally on the defensive when I receive questions about automated claims solutions.

Robots replacing humans? Computer says no? One size fits all? No thank you! What do our clients actually want? Good question!

Technology is clearly here to stay, and, throughout my career in the insurance market, it has been constantly bedded into my claims DNA that we need to evolve to stay at the top of our game. The challenge, then, is how to utilise technology while maintaining our core values and the vital human element, ensuring we don’t alienate our clients.

While researching “Claims AI,” I found a lot of enthusiasm for innovation but also some caution from industry experts. It was certainly a relief to read that efforts are being made to bring customers along on the technology journey. Bespoke solutions are being created for some very different first- and third-party product lines, and robots aren’t destined to take over the world!

While the last point remains uncertain, in my current company, we strive to excel daily and are committed to continuous improvement. What resonates with me is the ethos of doing the simple things well and embracing incremental change without wholesale revolution.

Let’s apply this to what is actually going on in the claims space:

• AI can expedite the claims process. For example, AI algorithms and image recognition swiftly analyse claim data, assess damages, and determine coverage using visual data from property inspections or accident scenes, including drone footage and automobile telematics.
• It can help detect fraud and predict trends.
• It provides a theoretically seamless continuum through from any automated underwriting stage.

Sounds great, but what about some real-life examples of how this actually benefits customers?

• Zurich’s “Zara” automated system analyses images, videos, and audio recordings, reducing processing time by over 30% and boosting customer satisfaction above 80%.
• Ernst & Young present a case study on real-time claim document processing, albeit with a current success rate of only 70%.
• State Farm offers the “Pocket Agent” chatbot for customer inquiries on underwriting and claims.

Seems promising, but what about deeper ethical concerns?

• Ensuring AI algorithms use accurate, unbiased data.
• Maximising outputs from vast unstructured (as opposed to structured) data that, until now, have gone largely neglected within the insurance industry.
• Mitigating against privacy concerns, like misuse of personal data.
• Striking a balance between technology investment, including skilled data scientists, without neglecting to develop other crucial human talent.

And this latter point is very much where my focus remains. I embrace the tangible value that AI can offer, particularly when it is not burdensome for customers and can enhance routine or specific parts of the claims handling process, like auto-triaging claims, document labelling and e-discovery tools.

However, I’m still primarily committed to preserving human-centricity in complex insurance operations. It’s not merely about interacting with customers on a personal level, although that’s certainly crucial. It’s also about providing critical oversight to ensure that transformation initiatives, such as AI integration, align with ethical and customer-centric principles.

Penning this article has proven to be cathartic in confronting my unconscious biases. It’s allowed me to recalibrate my perspective towards AI, recognising its potential, while staying true to my core values. After all, I’m a human being first and foremost, and I suspect my sentiments resonate with countless other insurance consumers worldwide who still prefer the reassurance of dealing with a real person rather than a computer that may say “no”!

Image: Provided by Mission Underwriters/Canva.

Guest Post This post was created just for Claims Media by a guest contributor.