Gap Analysis | Project Northern Lights

Components Complete

Require Migration

New Implementation

Total Components

✅

Existing Components (Ready to Migrate)

These components have working implementations in the current codebase

Component	Current Location	Status	Target
Workiva OAuth2 Client	`app/services/workiva.py`	✓ Complete	func_bridge
Workiva Read/Write Ops	`app/api/routes/workiva.py`	✓ Complete	func_bridge
Multi-Agent Orchestration	`app/services/agentic_extractor.py`	✓ Complete	func_reason
PDF/Excel Extraction	`app/services/pdf_extracter.py`	✓ Complete	func_ingest
Document Search Tools	`app/services/document_tools.py`	✓ Complete	func_ingest
Step Logger / Audit	`app/services/step_logger.py`	✓ Complete	All Functions

🔄

Components Requiring Migration

These need adaptation to work with Azure services

Component	Current State	Status	Required Changes	Priority
LLM Integration	Google Gemini 2.5 Flash	⇄ Migrate	Replace with Azure OpenAI GPT-5.1, update prompts	High
Notifications	SMTP Email (Gmail)	⇄ Migrate	Replace with Teams/Copilot webhooks	Medium
Configuration	`mapper.yaml` + .env	⇄ Migrate	Move to Azure Key Vault + App Configuration	High
API Runtime	FastAPI + Uvicorn	⇄ Migrate	Convert to Azure Functions triggers	High

➕

New Components (To Implement)

These components need to be built from scratch

Component	Description	Status	Priority
Azure Event Grid	Event routing for blob triggers	✗ New	High
Azure AI Search (RAG)	Vector DB for historical context	✗ New	High
VNET + Private Endpoints	Zero-egress network architecture	✗ New	High
Copilot Studio Agent	Teams-based command interface	✗ New	Medium
SharePoint Integration	Drop zone with event triggers	✗ New	Medium
Azure Document Intelligence	Enhanced PDF/form extraction	✗ New	Medium
Azure Key Vault	Secrets and credential management	✗ New	High
Monitoring Dashboard	Log Analytics + Application Insights	✗ New	Low

📁

Code Migration Mapping

File-to-function mapping for migration

🚀

func_ingest

Document Parsing Function

pdf_extracter.py → extraction/

document_tools.py → tools/

step_logger.py → logging/

+ Document Intelligence → NEW

🧠

func_reason

LLM Orchestration Function

agentic_extractor.py → agents/

Google Gemini → Azure OpenAI

+ AI Search RAG → NEW

+ Prompt Templates → NEW

🔗

func_bridge

Workiva API Bridge

workiva.py (service) → api/

workiva.py (routes) → handlers/

mapper.yaml → Key Vault

+ Retry Logic → ENHANCE

🗣

Copilot Studio

Teams Interface Agent

notifications.py → Teams Webhook

+ Agent Definition → NEW

+ Adaptive Cards → NEW

+ Power Automate → NEW

Implementation Timeline by Phase

Azure Infrastructure

VNET, Storage, Key Vault

Functions Migration

ingest, reason, bridge

Intelligence Layer

OpenAI, AI Search

Interface Layer

SharePoint, Copilot

Integration & Test

E2E Testing, Deploy

Effort Estimation Summary

Based on existing codebase analysis and target requirements

~4,000

Lines of Existing Code

~60%

Code Reuse Potential

Implementation Phases