A policy-checking workbench built around agents, for the people who sign

Why we built it

A broker’s promise to a client is the binder. What the client keeps is the policy. The work of making sure the two say the same thing — and that the renewal says what the client was told about the expiring — is line-by-line reading across four documents, each worded by a different carrier in its own way. It is exactly the work a language model is good at, as long as the comparison itself is deterministic and every difference is cited to the page.

So InsightProofing starts from engines and agents together. The engines read every document into one canonical coverage model and decide pass, fail and change; nine agents explain what that means, draft the carrier’s correction and the client’s summary, and prepare them for a person to sign.

What we hold ourselves to

• AI first, people decide. Agents reason with the model by default; nothing reaches a carrier or a client without a person.
• Every difference cited. Every pass and fail carries the page on both documents; a model sentence with a number the facts don’t hold is discarded.
• One field, however it is worded. A renamed label is never a dropped field.
• Honest demo data. The brokerage, carriers, insureds and documents on this site are invented, and labelled so. We do not claim customers we do not have.

The Insight broker suite

InsightDealDesk, InsightDueDiligence, InsightRollup, InsightProofing, InsightPlacement and InsightAdvocate share one platform: the same identity, the same agent framework and autonomy dial, the same Rules Studio and decision record. Company intelligence comes from InsightPulse 360 and document extraction from InsightXtract.

VerticalServe

InsightProofing is built by VerticalServe, part of the InsightLake enterprise AI suite. Our Innovation PODs deploy and extend it inside a broker’s own environment.