Trust & Evidence

FAQ

Straight answers to the questions buyers and the people they report to actually ask.

Does Velerian replace our adjusters or examiners?

No. Velerian is decision support, not an autonomous decision-maker. The agent does the heavy lifting and produces a documented recommendation; a qualified person on your team makes the final call and stays accountable for it.

What stops the AI from making something up?

Three things. Every output is checked by an automatic evaluator before it's released, and anything that doesn't pass is stopped. Every decision has to cite its actual sources — policy language, the claim record — not assert things from thin air. And if the agent can't support an answer, it stops rather than guessing ("fail-closed"). Nothing reaches a person unless it's backed by the record.

Where does our data live, and who can see it?

Velerian runs on AWS in U.S. regions, and your data stays in your environment. Access is scoped and every agent action is logged. You control data residency, encryption, and access posture. (Bring-your-own-AWS is on the roadmap.)

Are you SOC 2 certified?

Not yet — and we won't say otherwise. A SOC 2 Type I engagement is planned, with Type II on a 12–18 month arc. In the meantime, design partners get earned trust earlier through a documented evidence package and a joint security review.

Has this been validated with real carriers yet?

Our claims packs are production-ready — they've passed a multi-criterion governance and evidence gate. They are not yet design-partner-validated; we're selecting our first partners now, and we won't claim customer validation we don't have. That honesty is the point of this whole section.

How is this different from the AI already in our claims system?

The AI built into a core claims system is bundled with that platform. Velerian runs alongside your claims system — Guidewire, Duck Creek, EIS, Insurity, Majesco, or legacy — and takes the audit-bearing slice the embedded AI doesn't cover. You keep control of the governance and evidence trail independently of your core-system roadmap.

How is this different from RPA / automation?

RPA executes fixed rules and produces logs. A governed agent reasons over the specifics of a claim, cites its sources, and explains its decision — and produces an evidence trail, not just a log of steps taken.

What does it take to get started?

A pilot is scoped small: one use case, one business unit. Velerian reads from your claims system through supported integration patterns, and writes its output back as an evidence package to your claim notes or file. The integration footprint stays small by design.

What happens if the agent is wrong and we get sued?

Final authority always stays with your adjuster or examiner — Velerian doesn't make the decision for you. And because every decision ships with its reasoning, citations, and approval chain, your defense counsel has a clean, complete record of how the decision was reached.