LOB-Aware Submission Routing: Email Parsing, Product Detection, and Dynamic Document Checklists Across 16 Lines of Business
How InsightUW detects that an email about "slip and fall at a warehouse" is a General Liability submission, routes it to the right team, selects the right product, and generates the right document checklist — all before a human touches it.
The Problem
A mid-size commercial carrier writes 16 lines of business. Each line has different products, different required documents, different underwriting teams, and different appetite rules. When a broker email arrives at submissions@company.com, someone has to figure out:
- Which LOB? — Is "cyber extortion coverage for a hospital" Cyber, Healthcare, or both?
- Which product? — General Liability has 6 products: premises/ops, products/completed ops, contractual, professional, umbrella, excess. The broker said "GL" — which one?
- Which documents? — A Cyber submission needs a network security questionnaire; a Marine Cargo submission needs a bill of lading. Request the wrong checklist and you waste 3 days in back-and-forth.
- Which team? — The CML team handles GL and Property. The SER team handles Marine and Construction. Route to the wrong team and the submission sits unworked.
The result: 30% of submissions are misrouted on first touch. Each misroute adds 1-2 days to cycle time. Brokers notice.
The InsightUW LOB Detection Engine
InsightUW uses a three-layer detection system: email keyword matching, product cascade from LOB, and dynamic document checklist generation.
graph TD
subgraph Input["Broker Email"]
A["Subject: GL Submission —<br/>Pacific Warehouse LLC"]
B["Body: ...slip and fall exposure,<br/>forklift operations, loading docks..."]
C["Attachments: Acord 125.pdf,<br/>Loss_Runs.xlsx"]
end
subgraph Detect["LOB Detection"]
D["Keyword Scan<br/>Subject + Body"]
E["LOB Match<br/>GL (score 0.94)"]
F["Product Cascade<br/>GL → Premises/Operations"]
end
subgraph Route["Routing Decision"]
G["Team: CML<br/>(Commercial Lines)"]
H["Mailbox Override?<br/>No — auto-detected"]
I["Fifo Assignment<br/>→ Kevin Patel"]
end
subgraph Docs["Document Checklist"]
J["GL Required Docs:<br/>Acord 125, Acord 126,<br/>Loss Runs, Financial Stmts"]
K["Completeness: 50%<br/>2 of 4 received"]
end
A --> D
B --> D
D --> E
E --> F
F --> G
G --> H
H --> I
E --> J
C --> K
How Email Keywords Map to LOBs
InsightUW maintains a keyword dictionary per LOB. Each keyword carries a weight. The LOB with the highest cumulative score wins.
General Liability Keywords
| Keyword / Phrase | Weight | Example Context |
|---|---|---|
| "general liability", "GL" | 1.0 | "GL Submission — Pacific Warehouse" |
| "premises", "operations" | 0.8 | "slip and fall exposure at premises" |
| "bodily injury", "BI" | 0.7 | "BI/PD limits of $1M/$2M" |
| "property damage", "PD" | 0.7 | "PD exposure from forklift operations" |
| "ACORD 125", "ACORD 126" | 0.6 | Attachment filename detection |
| "products liability", "completed operations" | 0.5 | "products/completed ops coverage needed" |
| "contractual liability" | 0.5 | "blanket contractual coverage" |
| "additional insured" | 0.3 | "AI endorsements required for 12 tenants" |
Cyber Liability Keywords
| Keyword / Phrase | Weight | Example Context |
|---|---|---|
| "cyber", "cyber liability" | 1.0 | "Cyber Submission — NovaPay" |
| "data breach", "breach response" | 0.9 | "data breach coverage for PII" |
| "ransomware", "cyber extortion" | 0.8 | "ransomware sub-limit of $5M" |
| "network security" | 0.7 | "network security liability" |
| "privacy liability" | 0.7 | "privacy regulatory defense" |
| "SOC 2", "SOC2" | 0.5 | "SOC 2 Type II certified" |
| "PCI DSS", "PCI compliance" | 0.5 | "PCI DSS Level 1 merchant" |
| "MFA", "multi-factor" | 0.3 | "MFA deployed across 95% of endpoints" |
Marine Hull & Machinery Keywords
| Keyword / Phrase | Weight | Example Context |
|---|---|---|
| "marine hull", "hull & machinery", "H&M" | 1.0 | "Marine H&M Submission" |
| "vessel", "ship", "fleet" | 0.8 | "12-vessel fleet, Panamax class" |
| "gross tonnage", "GT", "DWT" | 0.7 | "82,400 GT bulk carrier" |
| "class society", "Lloyd's Register", "DNV" | 0.6 | "Classed by Lloyd's Register" |
| "flag state" | 0.6 | "Marshall Islands flag" |
| "P&I", "protection and indemnity" | 0.5 | "P&I covered separately by IG club" |
| "salvage", "general average" | 0.4 | "includes general average coverage" |
Medical Malpractice Keywords
| Keyword / Phrase | Weight | Example Context |
|---|---|---|
| "medical malpractice", "MedMal", "med mal" | 1.0 | "MedMal Submission — City General" |
| "professional liability — medical" | 0.9 | "physician professional liability" |
| "hospital", "healthcare facility" | 0.7 | "450-bed teaching hospital" |
| "physician", "surgeon", "specialist" | 0.6 | "32 employed surgeons" |
| "claims-made", "retroactive date" | 0.5 | "claims-made with 2018 retro date" |
| "tail coverage", "ERP" | 0.5 | "tail coverage for departing physicians" |
| "credentialing", "privileging" | 0.3 | "credentialing standards documented" |
Multi-LOB Conflict Resolution
When an email matches multiple LOBs, InsightUW uses a scoring algorithm:
graph LR
subgraph Email["Ambiguous Email"]
A["Subject: Cyber + E&O for<br/>Fin Tech Payments Corp"]
end
subgraph Scores["LOB Scores"]
B["Cyber: 3.2<br/>(cyber, data breach, Soc2)"]
C["E&O: 2.1<br/>(professional liability, fintech)"]
D["D&O: 0.8<br/>(officers mentioned)"]
end
subgraph Resolution["Resolution"]
E["Primary LOB: Cyber<br/>(highest score)"]
F["Secondary LOB: E&O<br/>(flagged for review)"]
end
A --> B
A --> C
A --> D
B --> E
C --> F
When the gap between the top two LOB scores is less than 1.0, InsightUW flags the submission for manual LOB confirmation by the Underwriting Assistant. The analyst sees: "Detected: Cyber (3.2) and E&O (2.1). Please confirm primary LOB."
Product Cascade: From LOB to Specific Product
Each LOB contains multiple products. Once the LOB is detected, InsightUW narrows to the specific product using secondary keyword analysis.
General Liability Products
| Product | Trigger Keywords | ACORD Forms |
|---|---|---|
| Premises/Operations | "premises", "slip and fall", "bodily injury at location" | ACORD 125, 126 |
| Products/Completed Ops | "products liability", "completed operations", "manufacturing" | ACORD 125, 126, 141 |
| Contractual Liability | "contractual", "hold harmless", "indemnification" | ACORD 125, 126 |
| Professional Liability | "professional services", "E&O", "errors and omissions" | ACORD 125, 126, professional app |
| Umbrella | "umbrella", "excess over GL", "follow-form" | ACORD 125, umbrella app |
| Excess | "excess", "high excess", "buffer layer" | ACORD 125, excess app |
Cyber Liability Products
| Product | Trigger Keywords | Forms |
|---|---|---|
| First-Party (Breach Response) | "breach response", "notification costs", "forensics" | Cyber App, Network Questionnaire |
| Third-Party (Liability) | "privacy liability", "regulatory", "PCI fines" | Cyber App, Network Questionnaire |
| Full Cyber (First + Third) | "full cyber", "comprehensive cyber" or both first & third keywords | Cyber App, Network Questionnaire, Supplemental |
| Tech E&O + Cyber | "tech E&O", "technology professional" | Tech E&O App, Cyber App |
Marine Hull & Machinery Products
| Product | Trigger Keywords | Forms |
|---|---|---|
| Hull & Machinery (ITC) | "H&M", "Institute Time Clauses", "hull all risks" | Marine H&M App, Vessel Schedule |
| Total Loss Only (TLO) | "total loss only", "TLO", "constructive total loss" | Marine TLO App |
| Increased Value (IV) | "increased value", "IV", "disbursements" | Marine IV App |
| War & Strikes | "war risks", "strikes", "war & strikes" | War Risk App |
| Loss of Hire | "loss of hire", "loss of earnings", "off-hire" | Loss of Hire App |
Medical Malpractice Products
| Product | Trigger Keywords | Forms |
|---|---|---|
| Hospital Professional Liability | "hospital", "facility", "institutional" | Hospital MedMal App, Claims History |
| Physician Professional Liability | "physician", "individual doctor", "surgeon" | Physician App, CV/Resume |
| Allied Health | "nurse practitioner", "PA", "allied health" | Allied Health App |
| Clinical Trial Liability | "clinical trial", "research", "IRB" | Clinical Trial App, Protocol Summary |
Dynamic Document Checklists: 4 LOBs Compared
When a submission is created, InsightUW generates a LOB-specific document checklist. Here are the full checklists for four LOBs:
General Liability Checklist
| # | Document | Category | Required | Auto-Detect Pattern |
|---|---|---|---|---|
| 1 | ACORD 125 (Commercial Insurance Application) | application | Yes | Filename: "ACORD125" or "commercialapp" |
| 2 | ACORD 126 (Commercial General Liability Section) | application | Yes | Filename: "ACORD126" |
| 3 | 5-Year Loss Runs (current carrier) | loss_runs | Yes | Filename: "lossrun" or "claimhistory" |
| 4 | 3-Year Audited Financial Statements | financials | Yes | Filename: "financial" or "audit" |
| 5 | Certificate of Insurance (current) | certificate | No | Filename: "COI" or "certificate" |
| 6 | Safety/Risk Management Program | supplemental | No | Filename: "safety" or "riskmgmt*" |
| 7 | Vehicle Schedule (if auto exposure) | schedule | Conditional | Filename: "vehicle" or "autoschedule*" |
| 8 | Subcontractor Certificates (if construction) | supplemental | Conditional | Filename: "subcontractor" |
Cyber Liability Checklist
| # | Document | Category | Required | Auto-Detect Pattern |
|---|---|---|---|---|
| 1 | Cyber Liability Application | application | Yes | Filename: "cyberapp*" |
| 2 | Network Security Questionnaire | supplemental | Yes | Filename: "networksecurity" or "ITquestionnaire" |
| 3 | SOC 2 Type II Report (or equivalent) | compliance | Yes | Filename: "SOC2" or "SOC2*" |
| 4 | 3-Year Cyber Loss History | loss_runs | Yes | Filename: "cyberloss" or "cyberclaim" |
| 5 | IT Infrastructure Diagram | supplemental | No | Filename: "networkdiagram" or "infrastructure*" |
| 6 | Incident Response Plan | supplemental | No | Filename: "incidentresponse" or "IRP*" |
| 7 | PCI DSS Compliance Certificate | compliance | Conditional | Filename: "PCI" |
| 8 | Privacy Policy (public-facing) | supplemental | No | Filename: "privacypolicy*" |
| 9 | Business Continuity Plan | supplemental | No | Filename: "BCP" or "businesscontinuity*" |
Marine Hull & Machinery Checklist
| # | Document | Category | Required | Auto-Detect Pattern |
|---|---|---|---|---|
| 1 | Marine H&M Application | application | Yes | Filename: "marineapp" or "hullapp" |
| 2 | Vessel Particulars / Schedule | schedule | Yes | Filename: "vessel" or "fleetschedule*" |
| 3 | Classification Society Report | survey | Yes | Filename: "classreport" or "survey*" |
| 4 | 5-Year Marine Loss History | loss_runs | Yes | Filename: "marineloss" or "hullclaim" |
| 5 | Condition Survey (vessels > 15 years) | survey | Conditional | Filename: "conditionsurvey*" |
| 6 | Flag State Certificate | certificate | Yes | Filename: "flagcert" or "registration*" |
| 7 | ISM Code Certificate (DOC/SMC) | compliance | Yes | Filename: "ISM" or "DOC" or "SMC" |
| 8 | ISPS Certificate | compliance | Yes | Filename: "ISPS" |
| 9 | P&I Club Confirmation | certificate | No | Filename: "P&I" or "PandI" |
| 10 | Trading Area Map / Schedule | supplemental | No | Filename: "tradingarea" or "voyage*" |
Medical Malpractice Checklist
| # | Document | Category | Required | Auto-Detect Pattern |
|---|---|---|---|---|
| 1 | Medical Malpractice Application | application | Yes | Filename: "medmalapp" or "malpracticeapp" |
| 2 | Physician Roster / Staffing Schedule | schedule | Yes | Filename: "physicianroster" or "staffschedule" |
| 3 | 10-Year Claims History (extended for MedMal) | loss_runs | Yes | Filename: "claimhistory" or "lossrun" |
| 4 | Risk Management Program | supplemental | Yes | Filename: "riskmanagement" or "patientsafety" |
| 5 | Credentialing & Privileging Policy | compliance | Yes | Filename: "credentialing" or "privileging" |
| 6 | Surgical Volume by Specialty | supplemental | Yes | Filename: "surgicalvolume" or "procedurecount" |
| 7 | Financial Statements (3-year) | financials | Yes | Filename: "financial" |
| 8 | Joint Commission Accreditation Letter | compliance | Conditional | Filename: "JCAHO" or "jointcommission*" |
| 9 | CME Compliance Report | compliance | No | Filename: "CME" or "continuingeducation*" |
| 10 | Telemedicine Protocol (if applicable) | supplemental | Conditional | Filename: "telemedicine" or "telehealth" |
| 11 | Prior Carrier Declination Letter (if applicable) | supplemental | Conditional | Filename: "declination" or "non-renewal" |
The Scenario
A Gallagher broker sends a single email to submissions@company.com covering four insureds:
From: sarah.martinez@gallagher.com
Subject: New submissions — 4 accounts attached
Body:
1. Pacific Warehouse LLC — GL, premises/ops, $1M/$2M limits. Warehouse and distribution. Slip and fall history.
2. CipherShield Technologies — Cyber, full first and third party, $10M limit. SaaS company, SOC 2 certified, 500K user records.
3. Nordic Bulk Carriers — Marine H&M, 8-vessel fleet, Panamax, total value $280M. Lloyd's Register classed.
4. Valley Health Medical Center — MedMal, 320-bed hospital, 28 surgeons, claims-made. Current carrier non-renewed.Attachments: GL_ACORD_125_Pacific.pdf, GL_Loss_Runs_Pacific.xlsx, Cyber_App_CipherShield.pdf, SOC2_CipherShield.pdf, Marine_App_Nordic.pdf, Vessel_Schedule_Nordic.xlsx, MedMal_App_Valley_Health.pdf, Claims_History_Valley_10yr.pdf
InsightUW Processing
InsightUW detects 4 distinct submissions within a single email:
sequenceDiagram
participant Broker as Gallagher Broker
participant Parser as Email Parser
participant LOB as LOB Detection Engine
participant Fifo as Fifo Assignment
participant CML as CML Team (GL)
participant CYB as Cyber Team
participant SER as Specialty Team (Marine)
participant MED as Healthcare Team (Med Mal)
Broker->>Parser: 1 email, 4 accounts, 8 attachments
Parser->>Parser: Detect 4 submission blocks<br/>by numbered list parsing
Parser->>LOB: Block 1: "GL, premises/ops,<br/>slip and fall"
LOB->>LOB: Score: GL=3.8, Property=0.4
LOB->>CML: SUB-2026-GL-0118<br/>→ Kevin Patel (Fifo)
Parser->>LOB: Block 2: "Cyber, SOC 2,<br/>500K user records"
LOB->>LOB: Score: Cyber=4.1, E&O=0.9
LOB->>CYB: SUB-2026-CY-0044<br/>→ Jennifer Walsh (Fifo)
Parser->>LOB: Block 3: "Marine H&M,<br/>8-vessel fleet, Panamax"
LOB->>LOB: Score: Marine=4.5, Cargo=0.3
LOB->>SER: SUB-2026-MH-0043<br/>→ James Mitchell (Fifo)
Parser->>LOB: Block 4: "Med Mal, hospital,<br/>claims-made"
LOB->>LOB: Score: Med Mal=4.2, GL=0.6
LOB->>MED: SUB-2026-MM-0029<br/>→ Patricia Morgan (Fifo)
Results in 12 Seconds
| # | Insured | Detected LOB | Product | Team | Assigned To | Docs Matched | Completeness |
|---|---|---|---|---|---|---|---|
| 1 | Pacific Warehouse LLC | General Liability | Premises/Operations | CML | Kevin Patel | 2 of 8 | 25% |
| 2 | CipherShield Technologies | Cyber Liability | Full Cyber (1st + 3rd) | CYB | Jennifer Walsh | 2 of 9 | 22% |
| 3 | Nordic Bulk Carriers | Marine H&M | Hull & Machinery (ITC) | SER | James Mitchell | 2 of 10 | 20% |
| 4 | Valley Health Medical Center | Medical Malpractice | Hospital Prof Liability | MED | Patricia Morgan | 2 of 11 | 18% |
Each Underwriting Assistant sees their submission in their work queue with the correct LOB-specific checklist, missing documents highlighted, and the broker's email linked for context.
Document Auto-Matching
InsightUW maps the 8 attachments to the 4 submissions by filename pattern:
| Attachment | Matched To | Submission | Category |
|---|---|---|---|
| GL_ACORD_125_Pacific.pdf | Pacific Warehouse | SUB-2026-GL-0118 | application |
| GL_Loss_Runs_Pacific.xlsx | Pacific Warehouse | SUB-2026-GL-0118 | loss_runs |
| Cyber_App_CipherShield.pdf | CipherShield | SUB-2026-CY-0044 | application |
| SOC2_CipherShield.pdf | CipherShield | SUB-2026-CY-0044 | compliance |
| Marine_App_Nordic.pdf | Nordic Bulk | SUB-2026-MH-0043 | application |
| Vessel_Schedule_Nordic.xlsx | Nordic Bulk | SUB-2026-MH-0043 | schedule |
| MedMal_App_Valley_Health.pdf | Valley Health | SUB-2026-MM-0029 | application |
| Claims_History_Valley_10yr.pdf | Valley Health | SUB-2026-MM-0029 | loss_runs |
The 16 Lines of Business
InsightUW supports routing across all 16 LOBs:
| # | LOB | Code | Default Team | Products | Required Docs |
|---|---|---|---|---|---|
| 1 | General Liability | GL | CML | 6 | 8 |
| 2 | Commercial Property | PROP | CML | 4 | 9 |
| 3 | Commercial Auto | AUTO | CML | 3 | 7 |
| 4 | Workers Compensation | WC | CML | 2 | 8 |
| 5 | Cyber Liability | CY | CYB | 4 | 9 |
| 6 | Directors & Officers | DO | SPE | 3 | 7 |
| 7 | Errors & Omissions | EO | SPE | 4 | 6 |
| 8 | Employment Practices | EPLI | SPE | 2 | 6 |
| 9 | Marine Hull & Machinery | MH | SER | 5 | 10 |
| 10 | Marine Cargo | MC | SER | 3 | 8 |
| 11 | Inland Marine | IM | SER | 3 | 7 |
| 12 | Medical Malpractice | MM | MED | 4 | 11 |
| 13 | Construction | CON | SER | 3 | 10 |
| 14 | Excess Casualty | XS | SPE | 2 | 7 |
| 15 | Environmental | ENV | SER | 3 | 9 |
| 16 | Surety / Bond | SUR | CML | 3 | 8 |
All keyword dictionaries, product cascades, and document checklists are managed through the Reference Data admin page — no code changes required to add a new LOB or modify an existing checklist.
Key Metrics
| Metric | Manual Routing | LOB-Aware Routing |
|---|---|---|
| LOB misroute rate | 30% | 2.1% (flagged for human confirm) |
| Time to identify LOB | 5-15 minutes (human read) | < 2 seconds (keyword engine) |
| Wrong document checklist sent | 25% of submissions | 0% (LOB-specific auto-generated) |
| Multi-submission emails split correctly | Rarely (treated as one) | 96% auto-split accuracy |
| Document auto-match accuracy | N/A (manual upload) | 94% (filename pattern matching) |
| Days added by misroute | 1-2 days per misroute | 0 days (correct team on first touch) |
| CBP time on triage per submission | 8-12 minutes | < 30 seconds (review AI decision) |
InsightUW's LOB-aware routing eliminates the guesswork from submission triage — right LOB, right product, right documents, right team, every time. See it in action.