Risk Reporting
Recurring HTML diagnostics on the loss characteristics of your portfolio. Frequency, severity, triangle runoff, segment trends. The data layer that supports dynamic pricing and feeds AVA.
Risk Reporting is a set of recurring HTML diagnostics on a non-life insurance portfolio’s loss characteristics — loss-ratio development by accident year and segment, frequency and severity decompositions, triangle runoff with residuals and stability metrics, reporting-delay diagnostics, and segment-level trend views. Produced from your claims data, delivered on your cadence (typically monthly), shipped as static HTML you can share across the organisation without dashboard infrastructure, and structured for drill-down from portfolio view into segment view into the development triangle that explains it.
It serves two roles. As a standalone product, it gives an insurer a continuous view of how the book is moving. As the data foundation for AVA and indirectly for Dynamic Pricing , it is the diagnostic layer that makes the rest of the product line possible.
What is in the report
A standard Risk Reporting cycle covers three families of output.
Underwriting and accident reporting
The headline view of how the book is performing.
- Loss frequency by underwriting month and accident year, at the portfolio level.
- Loss ratio by the same dimensions, with model-vs-actual overlay.
- Average claim decomposed alongside frequency.
- Separated views for new business and renewals. They are different problems and we report them as such.
Portfolio structure monitoring
Where claims are concentrating, and where the book is drifting.
- Frequency by claim count: counts, relative counts, actual loss ratio, modelled loss ratio.
- Frequency by bonus level, with the same view applied by bonus segment.
- Cuts by performance, bonus, payment frequency, separate operator, fuel type, region, and status.
- Each cut available for new business, renewals, or total.
Other diagnostics
The pieces that do not fit the headline categories but make the report defensible.
- Claim reports showing what is being paid, where, and how.
- Reporting-delay diagnostics. When claims arrive versus when they occurred.
- Model-vs-reality detail reports, with an automatic flag for segments where the gap between model and reality is statistically significant.
See what a report looks like
Two live charts of the kind we ship every cycle. Both are anonymized samples from real engagements. Each chart is fully interactive: hover for the underlying numbers, toggle series in the legend, zoom by dragging.
How claim frequency develops month by month across the portfolio. This is the headline diagnostic for spotting drift in loss characteristics before it shows up in the loss ratio.
How the age distribution of newly acquired policies is shifting over time. Where the book is growing, where it is shrinking, and how that compares with what the tariff is set up to handle.
How it gets delivered
- HTML, not PDF. Static HTML reports you can open, search, and share. No PDF flattening, no infrastructure to host.
- Excel companions where useful. All the values behind every chart are exported to Excel alongside the HTML for analysts who want to pull them apart.
- Monthly cadence by default. Other cadences are available, but monthly is what most production reserving and pricing teams want.
- Custom segments and slices on request. The standard report covers the dimensions above. Bespoke cuts are quick to add.
How it connects to the rest
Risk Reporting is built to do two jobs at once.
- Stand on its own, for insurers that want a continuous view of the portfolio and do not run the diagnostics internally.
- Feed AVA. Risk Reporting’s data layer is exactly the input AVA needs to find segments where the risk model is mispricing. Run the two together and the diagnostics turn into a model-improvement cycle, not just a reporting one.
Want to see a sample
Write to info@com-pass.cz . We can walk through a redacted sample report from a comparable insurer in the first call.