# Appetite (https://riska.io/platform/appetite) _Appetite_ ## Focus every underwriter on high-value work. Clean risks go straight through, knockouts stop on arrival, and referrals reach an underwriter with the reasoning attached. Bind more in-appetite business per underwriter, and keep the risks you would never write off the loss ratio. ## Appetite set for each product line. One model decides which business moves on its own, which stops at the door and which waits for an underwriter. ### Bind more without adding headcount when clean risks clear on their own. Clean risks go straight through, end to end, without an underwriter touching them. ### Decline out-of-appetite risks before anyone spends an hour on them. The broker hears back while there is still time to place the risk elsewhere. ### Spend underwriting time where it counts. The underwriter starts at the decision instead of the research, with the risk factors that referred it laid out. ### Use agents to process more accounts than ever before. Add in controls and human checkpoints where a decision needs to be made. ## Appetite that learns from your own book. Model appetite on past performance and on any attribute in the Data Mesh, and audit how each one shaped the result. - **Never make a stale decision.**: Write your guidelines without code and release them at once. Every new submission is scored against the current version. - **Score the way your underwriters think.**: Group factors into categories, weight them against each other and limit how far one weak input can drag down an otherwise strong risk. - **Use our ML or bring your own.**: Riska's AI attributes and your own scores sit in the same framework, so every factor is weighed and explained the same way. - **Know why every score moved, live.**: Records change as new information arrives, and appetite keeps up. See each factor's contribution and how the score got to where it is now. > "“The first week we ran it we found we were declining more habitational than anyone in the room believed.”" — Underwriting lead, Commercial property MGA · 22 states ## Score any attribute the way you underwrite it. Because every attribute sits on the Data Mesh, you can score appetite on any of them in plain language or with no-code tools, then report on the result. ### Tiers. Bands, each with a score. A five year loss ratio under 40 percent scores 100, over 100 percent scores 0, and where the bands fall in between is up to you. ### Curves. For quantities where a bucket edge would be arbitrary. Interpolate along a line, an S curve or a log curve, so total insured value moves the score smoothly instead of in steps. ### Static. Presence is the whole signal. Sprinklered scores 90, and a blank scores whatever you decided a blank is worth. ### Pass-through. The value already is a score. A wildfire rating that runs 0 to 100 goes in untouched. ## Referrals an underwriter can clear quickly. A referral is not a dead end. The underwriter gets everything needed to decide, and the platform acts the instant the answer changes. - **Every referral explained**: See which rules fired, the score that put the risk in range and the note you wrote on the rule, together on the record. - **What would change the outcome**: Each factor shows how far it moved the score, so the underwriter knows which answer is worth chasing before going back to the broker. - **Nothing missing at the decision**: Outstanding requirements sit on the case and hold the next step until each one is satisfied, waived or marked not applicable. - **Every option side by side**: When a risk could go more than one way, the options are ranked and recorded with the one taken and why not the others. - **Knockouts on the record**: Out of appetite, recorded against the rule that fired and the score that put it in range. Every decline is one you can explain. - **Act the instant appetite changes**: Trigger any workflow when a record moves in or out of appetite, from a task for the underwriter to an email to the broker. ## Defend every decision months later. Every result is recorded and filed, so a decision made today can still be explained, and improved, long after the file closes. ### Point in time decisioning. Open any scored version and see the final score with its breakdown. The decisions you make today stay explainable to an auditor, a reinsurer or your own team. ### Appetite as a feedback loop. Turn appetite into a living model. Compare good and bad business against the scores it was written on, and tighten the bands with what your own loss ratio teaches you. ### Book wide. One scoring model across lines means one view of the book. See how each product line, workspace, team and org is performing side by side. ## Underwriting, end to end. Appetite is one decision in a longer path. The rest of the platform carries the same record from intake through renewal. ## Bring the referral you worked last week. We will encode a few of your appetite rules and run a live submission through them, from straight through to referral.