Proof / Architecture

Salesforce-native. AWS-powered AI. Zero data lakes.

inflooens runs entirely within your Salesforce org. Luna's AI backend runs on AWS. Your data never leaves systems you control.

System Architecture

Five-layer architecture

Clean separation of concerns with enterprise-grade patterns at every layer.

1

Presentation Layer

Salesforce Lightning Experience with custom LWC components

Luna AI Drawer (LWC)Pipeline ViewsLead Management UIBI Dashboards
2

Application Layer

Apex business logic with governor-limit-safe patterns

Luna Service ClassesEncompass Sync EngineCredit Pull OrchestratorMessage Center
3

Data Layer

15+ custom objects with 2200+ fields, full audit trails

Loan__cLead__cCredit_Report__cCondition__cMessage__c
4

Integration Layer

Platform events, REST APIs, and webhook handlers

Encompass Sync EventsCredit Provider APIsSMS/Email GatewaysAWS Lambda Callouts
5

AI Layer

AWS-hosted AI backend with vector search and LLM

AWS Lambda FunctionsPinecone Vector DBClaude (Anthropic)OpenAI Embeddings
Data Flow

How Luna analyzes a loan

From drawer click to AI insight in under 3 seconds.

1

Loan Officer opens loan record

User clicks Luna drawer in Salesforce Lightning

2

Apex queries loan data

All related objects queried with CRUD/FLS enforcement

3

Data sent to AWS Lambda

Secure HTTPS callout with JWT authentication

4

AI analysis performed

Luna analyzes loan, detects anomalies, predicts approval

5

Results cached and returned

Response stored in Salesforce, displayed in drawer

Salesforce Native

Your data stays in Salesforce

inflooens doesn't replicate your data to external databases. Everything runs within your Salesforce org with full platform security enforced.

  • No external data lakes
    All loan data stays in Salesforce objects you control
  • CRUD/FLS enforced
    All queries respect Salesforce security model
  • Sharing mode enabled
    Users only see data they have access to
Data Residency
Salesforce Org
All loan data, leads, conditions, credit reports, messages, documents
AWS Lambda
Stateless AI processing only—no data persistence
Encompass Integration

Bidirectional real-time sync

Not a nightly batch job. Real-time, event-driven synchronization.

Bidirectional Real-Time Sync

Changes in Encompass automatically sync to Salesforce. Changes in Salesforce sync back. No batch jobs, no delays.

Field-Level Mapping

500+ fields mapped between systems. Configurable per-client. JSON-based transformation rules.

Conflict Resolution

Timestamp-based conflict resolution with audit trail. System of record configurable per field.

Event-Driven Architecture

Salesforce Platform Events trigger sync operations. No polling. No wasted API calls.

AI Infrastructure

AWS-powered AI backend

Enterprise-grade AI infrastructure with no data persistence.

AWS Lambda

Serverless Compute

Stateless functions that scale automatically. No servers to manage. Pay only for compute time used.

Pinecone

Vector Database

Stores guideline embeddings for semantic search. Sub-100ms query latency. Guidelines only—no customer data.

Claude + OpenAI

LLM Processing

Claude (Anthropic) for analysis and reasoning. OpenAI for embeddings. No training on customer data.

Want a deeper technical dive?

Our technical team can walk you through architecture details, data flow, and integration patterns.

Talk to Our Technical Team