Common questions

Does the tool approve claims automatically?

No. It prepares evidence and possible issues for human review. Authorized professionals make final decisions.

Why keep older processing runs?

They create an audit trail and allow teams to compare changes in models or review instructions. The default library still shows only the latest run so older results do not clutter normal work.

Can several people work on the same claim?

Yes. Ratings, decisions, and comments are stored with the reviewed issue, making the work visible to the team. Formal assignment and claim-level approval routing are recommended next steps.

What happens when AI reviewers disagree?

Disagreement is useful information. It should make the claim easier to prioritize for human review, not be silently averaged away.

How is “10×” measured?

It is a surge-capacity goal and scenario. A team should benchmark representative claims to measure actual reviewer time, throughput, agreement, and error rates before making a performance claim.

Where can I try it?

The current demonstration is available at Claims Processor. Access may require the demonstration password and acceptance of confidentiality terms.