BriteCore’s Model Context Protocol (MCP) Server architecture extends the BriteCore platform with a flexible, AI-ready integration layer. MCP servers expose BriteCore capabilities as structured tools that can be called by AI agents, workflow assistants, and other MCP-compatible clients.
This architecture gives insurers a practical way to connect intelligent automation to core insurance operations while preserving clear boundaries, consistent API behavior, and platform-aligned workflows.
What MCP servers provide
BriteCore MCP servers translate business capabilities into callable tools. Each tool performs a focused action, such as retrieving policy data, creating a claim, generating a document, validating an underwriting rule, or executing an approved analytics query.
The current MCP server architecture includes more than 100 tools across core property and casualty operations, including:
- Claims and FNOL
- Claim accounting
- Contacts
- Documents and attachments
- Document templates
- Lines of business
- Notes
- Policies
- Quoting and rating
- Rules and underwriting
- Tasks and workflow
- Reporting and analytics
Architecture overview
BriteCore MCP servers are organized as a Python-based monorepo with shared infrastructure and independently deployable server packages.
At the foundation is a shared bc-common library that provides common server behavior, API communication, configuration, logging, health checks, and transport support. Individual MCP servers use this common foundation to expose specific BriteCore domains as tools.
Each server follows a consistent pattern:
- The server starts using a shared entry point.
- Tool files are discovered automatically.
- Functions decorated as MCP tools are registered.
- The server exposes those tools through MCP transport.
- Tool calls communicate with BriteCore through supported API or gateway interfaces.
- Responses are returned in a consistent JSON format.
This design makes the architecture easy to extend. New tools can be added by placing a tool file in the appropriate server and registering the function with the MCP tool decorator.
Dynamic tool discovery
BriteCore MCP servers use dynamic tool discovery. When a server starts, it scans its tool directory and automatically registers available tools.
This approach reduces manual wiring and helps keep each server modular. Developers can add, update, or remove tools without maintaining a separate central registry. It also supports a clean separation between shared platform behavior and domain-specific business logic.
Transport options
BriteCore MCP servers support flexible transport modes for different deployment and development scenarios.
The default transport is stdio, which is commonly used by MCP-compatible desktop and local agent clients. The servers also support HTTP transport for stateless service communication, including deployment patterns that require HTTP routing, container orchestration, or remote agent access.
Servers also expose a health endpoint that can be used for readiness and liveness checks in containerized environments.
BriteCore API and gateway access
MCP tools communicate with BriteCore through two primary integration paths.
The first path uses BriteCore API calls for standard REST-based operations. This is used for many platform workflows, including policy retrieval, claims data, rules, lines, tasks, documents, and quote operations.
The second path uses BriteCore gateway calls for workflows that rely on established internal platform services. Examples include FNOL creation, deliverable generation, and file upload workflows.
By supporting both paths, MCP servers can expose a broad range of current platform capabilities while maintaining compatibility with existing BriteCore services.
Consistent tool responses
BriteCore MCP tools return a consistent JSON response structure. Each response indicates whether the tool call succeeded, includes returned data when available, and provides an error message when the operation fails.
This consistency helps agent clients handle results predictably across different business domains. It also makes MCP workflows easier to debug, monitor, and compose into larger automated processes.
Available MCP servers and tools
Claims
The claims MCP server supports First Notice of Loss and related claims workflows. Available tools include creating an FNOL, submitting a claim, assigning a policy to a claim, retrieving claim details, managing loss information, setting loss causes, selecting coverages, retrieving applicable catastrophes, and saving catastrophe associations.
These tools help agents guide users through structured claims intake and claim setup processes.
Claim Accounting
The claim accounting server supports financial activity on claim exposures. Tools include listing exposures, adding reserves, retrieving accounting summaries, reviewing exposure transactions, retrieving payment information, and making claim payments.
These capabilities support reserve management, payment workflows, and claim financial review.
Claims Dashboard Statistics
The claims dashboard statistics server provides analytics-oriented tools for claims operations. It can retrieve dashboard metrics such as active claim counts, claims by status, open tasks, task assignment volumes, claims flow, oldest open claims, and task resolution timing.
These tools are useful for operational dashboards, management reporting, and agent-assisted claim workload review.
Contacts
The contacts server provides tools for searching and retrieving contact information. It supports contact lookup by search terms and full contact retrieval by contact ID.
These tools help agents connect workflows to the correct insured, claimant, agent, payee, or related party.
Documents
The documents server supports document and attachment workflows. Tools include generating deliverables, retrieving available deliverables, retrieving attachments, downloading attachments, emailing attachments, uploading documents, and uploading photos.
These tools allow agents to work with BriteCore files, forms, and communications as part of claims, policy, or contact workflows.
Document Templates
The document templates server supports a complete document template pipeline. Tools include converting PDFs to DOCX, converting PDFs to images, inspecting DOCX files, repairing DOCX tables, replacing static content with template placeholders, uploading templates, retrieving templates, deleting templates, validating templates, rendering PDFs, retrieving BriteCore rendering context, comparing document images, and managing session files.
This server is designed for advanced template development and document automation workflows.
Lines
The lines server exposes tools for line-of-business configuration, policy types, effective dates, quoting questions, and rate chain management. Tools can retrieve lines, states, policy types, effective dates, policy type items, quoting questions, and can create or modify policy type configuration.
A specialized rate chain patching tool allows an agent to update a specific rate chain step without passing the entire rate chain through the client interaction.
Notes
The notes server supports creating and retrieving notes on BriteCore entities. Notes can be attached to claims, policies, contacts, and other supported records.
These tools help maintain audit trails and preserve context across agent-driven workflows.
Policies
The policies server provides policy retrieval tools by contact and date range. Policy responses can include risks, contacts, billing information, claims, and photos.
Because policy payloads can be large, the server supports pagination and encourages focused retrieval.
Quoting
The quote server supports quote creation, rating, payload construction, coverage matching, driver updates, and vehicle updates.
Tools can create and rate quotes, rate existing quotes, build quote payloads, match extracted coverage names to BriteCore coverage schema, confirm fuzzy coverage matches, add drivers, list vehicles, and update vehicle details.
The quote workflow includes advanced support for coverage limit correction and re-rating when platform defaults differ from requested coverage values.
Rules
The rules server supports underwriting rules, straight-through processing, rule validation, rule configuration, and rule documentation. Tools can list rules, retrieve rules, create rules, update rules, validate rule structures, retrieve available triggers and effects, retrieve advanced trigger documentation, and create STP error codes.
These capabilities help agents assist with underwriting automation, rule configuration, and validation before deployment.
Tasks
The tasks server supports BriteCore workflow management. Tools can create tasks, modify tasks, retrieve tasks, retrieve a single task, retrieve task definitions, retrieve task history, and retrieve available filter options.
These tools help agents coordinate claim and policy work across users, queues, due dates, and task definitions.
Reporting and Analytics
The reporting server provides controlled access to analytics views, stored procedures, report categories, view documentation, coverage schema details, report examples, and validated SQL execution.
It supports report generation and data exploration while limiting execution to approved query patterns.
Policy Insights
The policy insights server is a lightweight framework server that can be used as a starting point for future policy-focused MCP capabilities.
Key technical patterns
Modular server design
Each domain has its own MCP server. This keeps tools organized by business capability and allows teams to deploy, test, and extend servers independently.
Shared platform foundation
Common behavior such as server startup, tool loading, BriteCore communication, logging, and health checks is centralized in the shared library. This reduces duplication and promotes consistent implementation across servers.
Async execution
MCP tools use asynchronous execution patterns for non-blocking API communication. This supports responsive agent workflows and efficient interaction with BriteCore services.
Fail-fast startup
Servers are designed to fail during startup if a tool cannot be loaded successfully. This prevents partially loaded servers from running with missing or broken capabilities.
Agent session awareness
Some workflows can use agent session context to create traceability across multi-step interactions. For example, claim creation can attach session-aware notes to support auditability and workflow review.
Server-side payload management
Several tools are designed to reduce the amount of complex payload data that must pass through an AI interaction. For example, tools can retrieve existing platform data, modify a targeted element, and submit the complete updated structure back to BriteCore.
This pattern improves reliability for complex workflows such as rate chain updates and quote coverage adjustments.
Configuration
MCP servers are configured through environment variables. Common configuration includes the BriteCore API base URL, API keys, SSL verification settings, transport mode, host, and port.
Database-backed reporting tools use database connection settings and a controlled list of allowed procedures.
This configuration model supports local development, containerized deployment, and environment-specific runtime behavior.
Common use cases
BriteCore MCP servers can support many agent-assisted workflows, including:
- Guided FNOL intake
- Claim status review
- Claim payment preparation
- Policy lookup
- Contact lookup
- Quote creation and rating
- Coverage mapping
- Underwriting rule configuration
- STP validation
- Task management
- Document generation
- Attachment handling
- Template development
- Operational reporting
- Analytics exploration
Benefits
BriteCore’s MCP Server architecture creates a structured bridge between AI agents and core insurance operations. It enables automation without requiring agents to work directly against unstructured systems or undocumented workflows.
Key benefits include:
- Clear tool boundaries
- Consistent response handling
- Modular domain ownership
- Flexible transport options
- Reusable shared infrastructure
- Support for both modern API and established gateway workflows
- Easier extension of agent-accessible capabilities
- Improved traceability for multi-step workflows
Summary
BriteCore MCP servers provide an open, extensible architecture for connecting AI-enabled tools to BriteCore’s insurance platform. By exposing core capabilities as structured MCP tools, BriteCore enables guided workflows, intelligent automation, and efficient access to platform services across claims, policies, quoting, underwriting, documents, tasks, and analytics.