Editor’s Note
Three announcements caught my attention this week. One concerns the submissions arriving at a managing general agent. Another concerns claims estimates. The third concerns how a carrier decides which brokers and submissions deserve more attention.
Each puts artificial intelligence inside an existing insurance workflow. The immediate promise is familiar: less time gathering information and more time making decisions. The more interesting question is what happens around those decisions. Who checks that the information is right? Who sets the priorities? What does the record show when someone later asks why a submission moved forward or an estimate changed?
The products are new, and their promised results still need to be tested in practice. Together, they offer a useful look at where AI is being put to work now.
– James W. Moore, Editor-in-Chief
Vertafore Moves AI to the Start of MGA Underwriting
Vertafore has introduced a Submission Processing Agent for managing general agents (MGAs) and specialty insurance workflows. It reads incoming emails, PDFs, and attachments, extracts information, creates a structured submission, and checks whether required details are missing. If they are, the system can request more information from the submitting agent before the file moves forward.
This targets work that can consume considerable time before an underwriter evaluates the risk. It could also give the underwriter a more usable starting point.
Why it matters: Submission intake shapes everything that follows. A missing field can delay a decision; a field extracted incorrectly can affect one. As more of this preparation moves to AI, MGAs and their carrier partners will need a way to check the extracted information against the original submission, particularly when it influences clearance, appetite, or authority.
NAMIC and Xceedance Offer Virtual Claims Estimating
The National Association of Mutual Insurance Companies (NAMIC) and Xceedance have launched a virtual estimating program for NAMIC members covering residential property, auto physical damage, and farm equipment claims.
The service combines AI-supported scope development, repair identification, document review, and quality checks with estimates validated by Xceedance-employed adjusters against carrier requirements. The organizations say the program is designed to deliver estimates in less than 48 hours. That is a stated service goal, not an independently established result.
Why it matters: The human review step is consequential. An estimate affects claim cost and the policyholder’s experience, and a faster first estimate is useful only if the scope and repair details hold up. Carriers considering this model should pay close attention to how an estimator resolves a disagreement with the AI output and how that change appears in the claim record.
Sixfold Brings Broker Strategy to the Underwriter’s Desk
Sixfold has launched Distribution Intelligence, a capability that evaluates brokers using signals including submission volume, completeness, appetite fit, and hit ratio. Underwriting leaders can set targets, such as growth in a particular segment. Sixfold then uses the broker and submission data to recommend actions within its underwriting workflow, including which in-appetite submissions might receive more attention.
There is a practical appeal here. Carriers have long tried to understand which distribution relationships produce business that fits the book they want to write. Connecting that analysis to the underwriter’s daily work could make it more useful.
Why it matters: These recommendations can influence which opportunities receive attention before a quote is made. Their value will depend partly on whether the system distinguishes a broker’s performance from the carrier’s own response. A low hit ratio, for example, may reflect poor submissions, slow turnaround, uncompetitive pricing, or some combination of the three. That distinction matters when a recommendation changes how the carrier treats a broker.
