How we helped an insurance underwriting group add $1.96M in net recurring profit.

Every new program brought another version of the same work.

A commercial insurance underwriting group was adding programs, but each team still handled submissions and carrier reports in its own way. It operated as a managing general agent, or MGA, which meant its underwriters made insurance decisions within authority granted by carriers.

Before an underwriter could evaluate a risk, someone had to read the broker's email and attachments, enter the details, check whether the submission fit the program, look for duplicates, and chase missing information. Once a policy was bound, staff had to assemble the carrier's monthly report from those same details.

Because each team had its own process, growth kept adding more versions of that work. Their underwriting leader needed capacity for eligible risks, but preparation and clearance were taking up that time. The leader funding the work also needed growth to produce lasting profit, which meant we had to improve the work around underwriting as well as measure the financial result.

40% of submissions didn't fit, but still received a full read.

We needed to see where the work was consuming their teams' time. And so we started with a one-week study in which underwriters, assistants, operations staff, and reporting staff logged their work. We compared those logs with twelve months of submission-inbox and policy-system records to see both the staff effort and the path submissions took.

The study found that underwriters spent ~33% of their week on intake, clearance, and missing-information follow-up. It also showed that 40% of submissions didn't fit the program but still received a full manual read. Submissions they could actually write were competing for attention with risks they couldn't.

The repeated work continued after the underwriting decision. Monthly carrier reporting took 40 staff-hours per program, even though intake had already required much of the same information. That connection shaped our build: we needed one submission record that both intake and reporting could use.

We made one submission record serve both ends of the process.

The teams recorded information in different ways, so we first mapped their existing fields to an agreed set. We built a shared record that followed each risk from the broker's email through the policy and carrier report.

With that record in place, we built an intake assistant to read incoming emails and attachments, fill in the details, and check the risk against program rules. When information was missing, the assistant flagged it and drafted a request back to the broker. It also helped prioritize submissions that fit the program, so underwriters could focus their attention on risks they could actually write.

We built the carrier-report builder around the same record because the duplication didn't end at intake. As a result, staff could use the information already captured for the policy rather than retype it across systems for the monthly report.

The workflow ran in a private environment, and their risk team took ownership of the rules we defined for AI use. AI prepared and checked information, but their underwriters kept every underwriting decision within their carrier-granted authority.

By connecting intake and carrier reporting, we removed repeated entry at both ends of the process while keeping the underwriting decision with their team.

We had to cut preparation time without losing agreement with their underwriters.

Faster preparation wouldn't help if the AI's work stopped matching their underwriting judgment. So their underwriters selected past submissions and marked the decisions we'd use as references. We tested the workflow on submissions we'd held out of development.

The pilot covered one program group and had two requirements: at least 33% less non-underwriting time per submission, with no drop in agreement with their underwriters' marked decisions. Both had to pass before the workflow could move into daily use.

The pilot passed both requirements, so we helped the participating teams put the workflow into daily use and retire repeated entry across their systems. We monitored usage and accuracy each week to keep track of how the workflow performed in their actual work.

Each program needed 30 fewer staff-hours of reporting work every month.

Monthly carrier reporting fell from 40 staff-hours to 10 per program, a 75% reduction. That freed 30 staff-hours from the same recurring task each month, giving their teams back time they'd been spending assembling information they already held.

Their finance team also independently measured $1.96M in added net recurring profit. That gave the leader funding the work the result they needed: growth that left the business with more recurring profit.

Their underwriting leader had a different concern: whether the teams could spend less time around each submission without losing the judgment the business relied on. The pilot answered that with less preparation time and no drop in agreement with their underwriters' decisions. Together, those results addressed both the capacity problem and the demand for lasting profit.

The group also recorded $16M more bound premium from incoming submissions and $2.4M in additional commission. That commission was a separate result from the recurring profit increase.

New programs could use the process we built.

The group was still growing, so the new workflow had to give their teams a way to add programs without creating another disconnected process. We trained their shared-services and data teams to add programs and operating companies to the shared intake and reporting workflow.

That workflow became the standard for new programs. Our handover gave the teams responsible for maintaining it the training to extend it, while their risk team kept ownership of AI-use rules and their underwriters kept decision authority.

See how we tested and measured the work.

The technical companion explains the shared record, historical testing, decision controls, and measurement definitions.

They could grow without rebuilding the work around every new program.

They wanted more underwriting business, but each new program was adding another version of the same preparation and reporting work. We measured that burden, connected both tasks through one submission record, and proved the workflow against their underwriters' decisions.

As a result, their teams had less routine work around each decision and a shared process for the programs that came next. Reporting fell from 40 to 10 staff-hours per program per month, and finance confirmed $1.96M in added net recurring profit. We helped them turn growth into more profit while keeping underwriting decisions with the team they trusted.