Auto Insights can help insurance teams automate claims performance analysis, drill into claims trends and spot anomalous claim rates. You can read more here on our website.
This article will cover:
Example insights from this use case
Recommended data structure for this use case
What sort of insights can Auto Insights help me uncover?
We've outlined some example questions which Auto Insights can help answer through a combination of its proactive insights, Missions and What caused this? analysis:
Automate claims performance analysis
Number of Claims by Department
Average Time to Payment by Agent
Number of Claims by Policy Type
Drill into claims trends
Trend of Claims by Claim Reason in Sydney
Seasonality of Claims
Compare Claim Amounts for Theft in Metropolitan and Regional areas
Average Time to Close Claims by Department and Claim Type
Spot anomalous claim rates
Average Number of Claims per customer
Claim Amount by Claim Reason in Sydney
Top 20 Claims by Claim Amount and Claim Reason last quarter
How do I structure my data?
Auto Insights requires structured, transactional data, with at least 1 measure (e.g. Number of Claims) and 5 segments (e.g. Claim Type). In addition, we recommend at least 7 months of data (at monthly or daily granularity) so you can take full advantage of Auto Insights' Unexpected Changes feature.
Example data structure
Please click on the picture to download a larger copy of the image.
Please refer to this article for more information about data structure.
Here are some of the typical segments we find in revenue data. A segment is a qualitative value, like names or categories:
Claim attributes: Claim reason, Claim type, Team, Department etc.
Policy attributes: Policy type, Policy name, Months since last claim, Agent assigned, claim channel, coverage type etc.
Claimant attributes:, Age, Gender, Location, etc.
A measure is a quantitative, numeric value. Some of the typical measures include customer ID, number of claims, claim amount, time to close claims, time to payment etc.
If you have any questions, please contact us at email@example.com.