An MGA Partnered With Notch to Automate Submission Intake and Underwriting Operations With Agentic AI
Achieving scalable intake automation across 10 P&C lines and 25 US states
A fast-growing US-based MGA was managing submission intake across 10 different property and casualty lines, operating in 25 states. As the business scaled, the team faced a familiar challenge: submission volume was increasing, but intake and underwriting operations were still heavily dependent on manual processing.
Submissions were arriving through multiple channels: broker emails, scanned documents, ACORD forms, attachments, faxes, and phone calls. Each channel created a different operational burden. Information had to be reviewed, extracted, checked against appetite, validated against state-specific rules, and entered into the policy administration system. The process required significant human effort, slowed down quote turnaround, and made it difficult to maintain consistency across lines, states, and underwriting teams.
The MGA partnered with Notch to automate its submission intake and triage workflow using agentic AI purpose-built for insurance operations.
The challenge: fragmented submissions, manual review, and state-by-state complexity
Before Notch, the MGA’s intake process relied on teams manually reviewing every incoming submission. A single submission could include an ACORD form, broker email, loss runs, supplemental applications, PDFs, scanned documents, and follow-up notes from phone calls. Critical information was often spread across multiple files and formats.
This created several operational challenges:
Information was fragmented across channels and documents. Underwriters and intake teams had to search manually for named insured details, exposures, requested coverage, limits, locations, prior carrier information, loss history, and missing data.
State-by-state rules added complexity. Because the MGA operated across 25 states, intake and triage could not rely on a single generic workflow. Business rules, eligibility logic, appetite criteria, and routing rules needed to adjust by state, line of business, and submission type.
Manual processing slowed down quote turnaround. Even when a submission was within appetite, teams spent valuable time preparing the file before underwriting could begin.
Incomplete submissions created back-and-forth. Missing information often surfaced late in the process, requiring additional broker communication and delaying first-pass review.
The PAS depended on clean structured data. The MGA’s policy administration system could only operate efficiently if the incoming submission data was accurate, structured, and complete.
The MGA needed more than document extraction. It needed a centralized AI operating layer that could understand submissions across formats, apply deterministic underwriting and operational rules, detect gaps, route work, and connect directly into existing systems.
The solution: one structured intake brain for submissions, rules, and routing
Notch implemented an agentic AI workflow for submission intake and triage. The system transformed unstructured submission inputs into one structured record that underwriting and operations teams could trust.
ACORDs, broker emails, scanned documents, attachments, faxes, and phone call inputs were ingested into a single AI-powered workflow. The system extracted key entities and exposures in one pass, including business details, locations, coverage requests, limits, class codes, prior policy information, and supporting documentation.
Instead of treating every file as a standalone document, Notch unified the submission into one structured view. The AI identified what was present, what was missing, what needed review, and where each answer came from.
When gaps or inconsistencies were detected, the system created source-linked flags. This allowed the MGA’s team to see not only that information was missing or questionable, but also which document, email, or field triggered the issue.
The workflow then applied deterministic rules, appetite logic, and state-by-state requirements. These rules helped the MGA automatically determine whether a submission should move forward, be prioritized, require clarification, or be routed to a specific underwriting queue.
The system also connected with the MGA’s policy administration system, enabling structured submission data to flow into existing operational infrastructure rather than creating another disconnected tool.
How the workflow worked
A broker submitted a new business opportunity through email, fax, scanned forms, or attachments.
Notch ingested the full submission package, including ACORDs, supplemental documents, broker notes, and supporting files.
The AI extracted entities, exposures, coverage details, and risk information into one structured submission record.
The system checked the submission against deterministic business rules, appetite logic, state-specific requirements, and line-of-business rules.
Missing, conflicting, or low-confidence information was flagged with source-linked evidence.
The submission was routed based on appetite, priority, completeness, state, line of business, and underwriting rules.
Clean structured data was passed into the MGA’s PAS, helping teams move faster from intake to quote.
The outcome: higher accuracy, faster processing, and stronger underwriting readiness
With Notch, the MGA automated a large portion of the submission intake and triage process across 10 P&C lines and 25 states.
The workflow achieved 99% accuracy in structured extraction and rule-based processing, helping the MGA reduce manual review and increase trust in automated intake decisions.
The MGA also saw more than 250% efficiency gains by reducing repetitive manual work, consolidating fragmented intake channels, and accelerating the preparation of underwriting-ready files.
In addition, the MGA improved first-pass submission completeness by automatically identifying missing data earlier in the process. This helped reduce broker back-and-forth and enabled faster quote turnaround.
Most importantly, the MGA created a scalable operating model for underwriting intake. Instead of relying on manual review across emails, PDFs, scans, faxes, and calls, the company now uses one AI-powered intake brain that can read, structure, validate, flag, route, and connect submission data directly into the systems its teams already use.
Core use case: Submission Intake & Triage
Notch turns ACORDs, broker emails, scanned documents, attachments, faxes, and call inputs into one structured submission record.
Entities and exposures are extracted in one pass. Missing or conflicting information is flagged with links back to the original source. Files are routed based on appetite, priority, state-specific rules, line-of-business logic, and underwriting workflow requirements.
The result is a faster, more consistent intake process that gives underwriting teams cleaner files, earlier visibility into gaps, and a stronger foundation for quote decisions.
Bottom Line - Business impact
99% accuracy across extraction and rules-based processing + More than 250% efficiency gains in submission intake operations.
By automating submission intake and triage across 10 P&C lines and 25 US states, Notch helped the MGA significantly reduce manual processing, improve submission readiness, and accelerate quote turnaround. The workflow achieved 99% accuracy across extraction and rules-based processing, delivered more than 250% efficiency gains in intake operations, and created a double-digit lift in first-pass completeness. With ACORDs, broker emails, scanned documents, attachments, faxes, and phone-based inputs unified into one structured workflow, the MGA was able to route files faster, reduce back-and-forth, connect clean data into its PAS, and build a scalable AI operating layer for underwriting operations.
Autonomous AI for operations leaders ready to turn complexity into advantage.
Deployed in weeks. Autonomous in months. Compounding for years.