What the showcase demonstrates
The showcase begins with the full claims network. Some clusters are already visible, but the real value comes from changing the structure of the data to match a specific line of enquiry.
One view focuses on connections between individuals. Instead of only showing the original claim records, it creates links where people share something in common. This helps analysts identify individuals who appear across otherwise separate claims. That may be normal for a doctor, but it is more unusual for a witness or policyholder to connect across multiple claims.
Another view adds location context by placing policyholders and garages on a map. Once we see distance, we can see that some people are travelling unusually far to use a particular garage. Why?
A further view reorganizes the data around repair behavior. In this example, one garage appears to handle a disproportionate amount of offside rear door repairs. That pattern does not prove fraud on its own, but it gives analysts a clear signal to investigate.
What you can evaluate
For product and engineering teams, this showcase demonstrates how the SDKs can support flexible fraud investigation workflows. It shows how the same underlying data can be reshaped around shared connections, geography or repair patterns, depending on what the analyst needs to test.
The workflow supports a core fraud analysis task: make connected claims data easier to interrogate so that analysts can spot patterns that may deserve closer investigation.
