What if you could simply ask your claims data a question?

Sedgwick is exploring new ways to use generative AI to help clients move from claims data to actionable intelligence faster, including a new capability called OmniAnalyst.

For years, claims analytics has helped organizations understand what is happening across their programs. Dashboards, reports and stewardship conversations can reveal trends, measure performance and help teams decide where to focus. But getting from data to an answer can still take work.

What if you could simply ask a question?

Why are indemnity costs increasing? Which states are driving the change? Where are we seeing patterns that deserve a closer look?

That is one of the ideas behind OmniAnalyst, a new AI-enabled capability Sedgwick is currently developing and testing with a small group of clients.

Turning questions into conversations

OmniAnalyst is being designed to allow users to interact with their claims data using natural language. Instead of knowing which report to run or how to manipulate the data, a user can ask a question and receive an analysis based on the information available to them.

And one question can lead naturally to another.

A user might start by asking how indemnity costs this year compare with last year. From there, they could explore which states are contributing to the change, look more closely at particular trends or continue asking questions as new insights emerge.

The goal is straightforward: make it easier to explore claims information and get to useful insights faster.

Building on the data advantage

The potential of AI depends heavily on the information behind it. That’s especially meaningful at Sedgwick, where the scale of our claims data gives us an extraordinary foundation from which to learn.

OmniAnalyst is being built to bring AI to that data, allowing users to explore information while maintaining established access controls. The capability is designed so users can only query the data they are authorized to access, while sensitive information, including HIPAA-protected and SPII data, is obfuscated or removed before it is made available to the AI.

Over time, we are exploring how additional sources of structured and unstructured information could make those conversations even richer.

The opportunity goes beyond answering questions faster. It is about helping people identify patterns, explore why something is happening and determine where a deeper conversation or intervention may be needed.

AI with people still firmly in the picture

As capabilities like OmniAnalyst evolve, the role of our data experts can evolve with them. Less time may be needed to retrieve information or produce routine analysis, creating more opportunity to help clients interpret what the information means, understand the context behind it and decide what to do next.

That same philosophy is guiding Sedgwick’s broader approach to AI. We are looking for places where technology can reduce friction, process information faster and surface insights sooner, while keeping people focused on the work that requires experience, judgment and empathy.

Building, testing, learning

OmniAnalyst is still in its early stages. We are currently testing the capability with a small group of Casualty clients, gathering feedback and continuing to refine the experience before a broader release.

That process matters.

With any application of AI, particularly one involving claims information, accuracy, governance and security have to be considered from the beginning. Human oversight remains important, and we continue to test how the capability responds to different questions and how its answers should be interpreted.

What we’re learning will help shape what OmniAnalyst ultimately becomes. And that may be the larger story.

AI in claims is quickly moving beyond ideas about what might someday be possible. Across Sedgwick, we’re putting new capabilities into real workflows, testing them against real business problems and learning where they can create meaningful value. 

OmniAnalyst is one more example of that work taking shape. Because the measure of innovation isn’t how much technology you can build. It’s whether that technology can help people see more clearly, make better decisions and ultimately deliver better outcomes.