cs.AISep 26, 2026

Governing the Edge: Automating Commercial Property and Casualty Insurance Underwriting via a Hybrid Local-Cloud Multi-Agent Framework

Authors: Vivek Kumar Singh, Gautam Bhowmick

Organizations: Independent Researcher McKinney, TX, USA · Independent Researcher Chicago, IL, USA

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Underwriters in commercial Property and Casualty (P&C) insurance spend 30 to 40% of their time on administrative work rather than risk judgment, and a single submission takes about 40 minutes by hand. We present Governing the Edge, a multi-agent framework for that layer, organized around a data-residency constraint: sensitive submission data must not leave the perimeter. Eleven agents and two deterministic control nodes, one a human-escalation interrupt, form a 13-node LangGraph workflow across two tiers. Agents touching raw submissions run locally on Gemma 2, each bound through a tool interface logging every invocation (synthetic stubs in the current prototype); only anonymized scores and non-identifying fields cross to Claude Sonnet in the cloud. Compliance rules are encoded as conditions on graph edges, so a non-compliant submission never reaches pricing. We walk through one full scenario end to end, a commercial auto submission whose principal driver carries serious violations, showing every agent call, every tool invocation, and the resulting routing decision. The framework processes a 40-minute manual submission in a few minutes on a single edge device, dominated by serialized on-device inference. On the hard-stop violation tier the framework enforces every rule correctly and reproducibly, since hard stops are deterministic predicates over fields extracted at temperature 0; overall compliance accuracy across the 20-scenario benchmark is 70%, with the remaining gap concentrated in softer, judgment-based tiers. It keeps a complete audit trail and respects the privacy boundary throughout. We release all code, compliance rules, tool stubs, and synthetic datasets.

Figures & tables

Fig. 1: End-to-end architecture of the Governing the Edge framework. Thirteen nodes execute across a LangGraph StateGraph with three conditional branch points and an explicit privacy boundary separating raw submission data (local tier) from anonymized score processing (cloud tier).
Fig. 2: Agent decision flow. Local agents (Gemma 2) invoke their bound tool interfaces and pass anonymized scores to cloud agents (Claude Sonnet) after the privacy boundary.
#AgentModelTierRole
1Doc ParserGemma 2LocalNormalize input
2IntakeGemma 2LocalExtract ACORD 125 (temp=0)
3aProperty/FleetGemma 2LocalFEMA/wildfire stubs
3bGeo/DriverGemma 2LocalMVR stub; driver risk
3cBusiness/OpsGemma 2LocalD&B/DOT stubs
3dRisk AggregatorClaude SonnetCloudMerge scores (anon.)
TABLE I: Agent Inventory (two control nodes omitted; total = 13).
IDLineRuleSev.Agent
CP-01Prop.Frame const. >> 3 storiesSTOP3a
CP-02Prop.TIV >> USD 10MSTOP3a
CP-03Prop.Habitational >> 4 unitsSTOP3a
CP-04Prop.Loss ratio >> 70%WARN3a
CP-05Prop.Bldg >> 40yr, no renoWARN3a
CA-01AutoFleet >> 20 vehiclesSTOP3a
TABLE II: P&C Compliance Rules Engine: 10 Validated Underwriting Rules
AgentTool FunctionProduction API
Document Parser Agentocr_document()AWS Textract / Azure Form Recognizer
Intake Agentvalidate_address() , lookup_sic_code()USPS Address Validation; D&B SIC/NAICS
Property Risklookup_fema_flood_zone() , lookup_wildfire_risk()FEMA Flood Map; CoreLogic Wildfire / Verisk
Geographic/Driverlookup_mvr()LexisNexis MVR / Equifax (consent required)
Business Risklookup_dnb_business() , lookup_dot_safety_rating()D&B Hoovers / Experian; FMCSA SAFER
Compliance Agentquery_rules_engine()Internal rules DB + state DOI filings
TABLE III: Tool Interface Map: Agent to Authoritative Production Source (synthetic stubs in prototype). Source acronyms: USPS, United States Postal Service; NAICS, North American Industry Classification System; FMCSA, Federal Motor Carrier Safety Administration.
CategoryTrigger
TIMEOUTLLM call exceeded its timeout.
LLM_API_ERRORAuthentication, rate-limit, or quota error from the model API.
LLM_PARSE_ERRORJSON/output parsing failed (also the catch-all default).
SCHEMA_VALIDATIONOutput failed Pydantic schema validation.
TOOL_FAILURETool/function-call invocation failed.
INCONSISTENCYOutput inconsistency or contradiction detected.
TABLE IV: Per-Agent Error Taxonomy: The Eight Failure Categories Tracked
Underwriting TaskManual TimeFramework (median)Tool / Agent
Document intake & field extraction16–24 min63.9 sDoc Parser + Intake
Driver/geographic risk (incl. MVR)8–12 min71.7 sGeographic/Driver ( lookup_mvr() )
Property/fleet exposure check5–10 min80.3 sProperty/Fleet ( lookup_fema_flood_zone() )
Business credit / DOT lookup10–18 min110.3 sBusiness/Ops ( lookup_dnb_business() , lookup_dot_safety_rating() )
Compliance rules check (10 rules)15–25 min39.2 sCompliance ( query_rules_engine() )
Coverage adequacy analysis10–15 min46.8 sCoverage Gap (cloud)
TABLE V: Per-Agent Latency on the Released 20-Scenario Run (M1, serialized on-premise inference). Manual times are from the McKinsey commercial-underwriting benchmark [ 9 ] ; framework times are median wall-clock per agent from the shipped telemetry ( evaluation/error_report.json ).
SystemCompliance Acc.CompletionAvg Latency
Monolithic Sonnet75.0%100%6.4 s
Linear Pipeline30.0%100%30.4 s
Governing the Edge70.0%100%268.8 s
TABLE VI: Baseline Comparison on 20 Synthetic Scenarios
Scenario TierMonolithicLinearGov. the Edge
Straightforward approval (4)4/40/44/4
WARNING compliance (6)3/60/62/6
Hard-stop violation (6)5/66/66/6
Low-confidence escalation (4)3/40/42/4
Overall75%30%70%
TABLE VII: Compliance Accuracy by Scenario Tier (committed 20-scenario run). Governing the Edge is the only system that enforces every hard stop (6/6) without collapsing on the other tiers.
More in this topic

Business Process Automation

Business Process Automation uses software, process models, and AI agents to execute, monitor, and improve organizational workflows.

See all 39 papers

Explore similar work

CardsList
Jul 14, 2026

AI-Native Insurance for Agentic AI: Pricing, Underwriting, and End-to-End Automation

Agentic AI introduces new insurance challenges because autonomous AI systems can make decisions, invoke tools, modify external environments, and interact with third-party services. This paper develops an AI-native mathematical framework for underwriting, pricing, and contract design for agentic AI deployments. A deployment is represented by a risk state that captures autonomy level, operational authority, permission exposure, governance maturity, and dependency concentration. The framework maps the risk state to event probabilities, loss severities, governance costs, premiums, deductibles, coverage allocation, and policy covenants, and formulates an optimization problem for insurance contract design under participation, profitability, and incentive compatibility constraints. The paper establishes structural properties of insurability, including characterization of an insurability region, monotone deterioration of feasibility with increasing exposure, and governance certification thresholds. Insurance is further interpreted as both an operational cost and a regulatory mechanism for AI deployment. A healthcare case study illustrates contract optimization, sensitivity analysis, and automated claims processing for agentic AI systems.

Quanyan Zhu

Department of Electrical and Computer Engineering and the NYU Center for Cybersecurity, NYU Tandon School of Engineering, New York University, Brooklyn, NY 11201, USA

AI Risk ManagementAI Agent GovernanceAgentic AI
AskPDF
Jul 8, 2026

Agentic AI and Retrieval-Augmented Models in Straight-Through Underwriting

Artificial intelligence (AI) is beginning to reshape actuarial practice, particularly in domains that require reasoning over unstructured documents, heterogeneous data sources, and regulated decision workflows. Actuaries now face a design space that ranges from traditional rule-based automation to large language models (LLMs), retrieval-augmented generation (RAG), and multi-agent agentic'' systems that plan, retrieve, call tools, and reflect. This paper examines how these emerging architectures can support actuarial priorities such as transparency, auditability, and human-in-the-loop governance, with a focus on straight-through decision processes. To make these ideas concrete, we develop and analyze an agentic AI framework for straight-through underwriting of small commercial Business Owner Policies (BOPs). We construct a synthetic but realistic experimental environment and compare three underwriting pipelines: (i) a single-LLM baseline, (ii) a naive RAG system, and (iii) a multi-agent Agentic RAG'' pipeline that combines targeted retrieval, third-party data checks, and explicit multi-step rule evaluation. The agentic system performs best overall, with the largest gains in multi-step and missing-information scenarios, where structured retrieval and reflection help the model avoid unsupported straight-through decisions.

Robert Richardson, Josh Meyers, Brian Hartman +1

Brigham Young University

Multi-Agent LLM SystemsAI-Assisted Decision MakingAgentic RAG
AskPDF
Jan 21, 2026

Agentic AI for Commercial Insurance Underwriting with Adversarial Self-Critique

Commercial insurance underwriting is a labor-intensive process that requires manual review of extensive documentation to assess risk and determine policy pricing. While AI offers substantial efficiency improvements, existing solutions lack comprehensive reasoning and internal mechanisms to ensure reliability in regulated, high-stakes environments. Full automation remains impractical and inadvisable when human judgment and accountability are critical. This study presents a decision-negative, human-in-the-loop agentic system that incorporates an adversarial self-critique mechanism as a bounded safety architecture for regulated underwriting workflows. In this system, a critic agent challenges the primary agent's conclusions prior to submitting recommendations to human reviewers. This internal system of checks and balances addresses a critical gap in AI safety for regulated workflows. Additionally, the research develops a formal taxonomy of failure modes to characterize potential errors by decision-negative agents. This taxonomy provides a structured framework for risk identification and management in high-stakes applications. Experimental evaluation using 500 expert-validated underwriting cases demonstrates that the adversarial critique mechanism reduces AI hallucination rates from 11.3% to 3.8% and increases decision accuracy from 92% to 96%. At the same time, the framework enforces strict human authority over all binding decisions by design. These findings indicate that adversarial self-critique supports safer AI deployment in regulated domains and offers a model for responsible integration where human oversight is indispensable.

Joyjit Roy, Samaresh Kumar Singh

Austin, Texas, USA · Leander, Texas, USA

LLM Self-CorrectionAI Agent SafetyHuman Oversight of AI
AskPDF