GenAI Cloud Service Request and Governance Management System

This repository supports a university assignment for 42904 Cloud Computing and Software as a Service. The project turns a previous research report on generative AI in cloud computing into a practical prototype for managing enterprise GenAI cloud service requests, reviews, approvals, and deployment reminders.

The assessed platform is Salesforce / Force.com PaaS. The main deliverable is a Salesforce Lightning prototype built with custom objects, relationships, validation rules, security controls, and flows. The Python service in this repository is only an auxiliary extension that demonstrates optional AI-assisted risk scoring. It is not intended to replace the Force.com solution and it does not depend on any paid cloud AI APIs.

Why Salesforce Is Primary

The assignment is assessed mainly on the Force.com prototype, report, and presentation. Salesforce is the primary platform because it directly supports the required:

• Lightning App
• Custom Objects and Fields
• Master-Detail and Lookup Relationships
• Validation Rules
• Custom Profiles and Tab Security
• Organisation-Wide Defaults
• Salesforce Flow for notifications, tasks, and scheduled reminders

The Python FastAPI service exists to support the governance theme by offering a lightweight local risk scoring API. Teams can demonstrate it as a supporting feature, but the main implementation work should remain in Salesforce.

Relationship to the Previous Report

The previous assignment explored cloud-based GenAI, AI-as-a-Service, privacy risks, edge-cloud architectures, cost concerns, governance, vendor lock-in, compliance, and the proposed PEACE-GenAI privacy-aware framework. This repository translates those ideas into a business process prototype where internal stakeholders request GenAI services and governance staff review the privacy, cost, compliance, and security implications before approval and deployment.

Repository Highlights

• force-app/ contains Salesforce DX-style scaffolding and object metadata for the required custom objects.
• docs/ contains the step-by-step manual implementation guide, report outline, security notes, workflow design, and demo script.
• src/genai_governance_api/ contains the auxiliary FastAPI service for local rule-based risk scoring.
• tests/ contains pytest coverage for the API health check and risk scoring rules.
• sample_data/ contains realistic CSV records for manual data entry or guided import.
• diagrams/ contains Mermaid diagrams for the report and presentation.

Install With uv

1. Install uv if it is not already installed.
2. From the repository root, run:

uv sync

Run The Auxiliary API

Start the local FastAPI service with:

uv run uvicorn genai_governance_api.main:app --reload

The API will expose:

• GET /health
• POST /risk-score

Run Tests

uv run pytest

Lint The Python Code

uv run ruff check .

Sample Risk Score Request

Use this sample request body against POST /risk-score:

{
  "use_case_type": "Document Analysis",
  "data_sensitivity": "High",
  "estimated_monthly_cost": 3500,
  "business_unit_budget": 5000,
  "cloud_provider": "Azure",
  "data_residency_region": "Australia",
  "supports_private_data": true
}

Example curl command:

curl -X POST "http://127.0.0.1:8000/risk-score" \
  -H "Content-Type: application/json" \
  -d "{\"use_case_type\":\"Document Analysis\",\"data_sensitivity\":\"High\",\"estimated_monthly_cost\":3500,\"business_unit_budget\":5000,\"cloud_provider\":\"Azure\",\"data_residency_region\":\"Australia\",\"supports_private_data\":true}"

Manual Salesforce Prototype Implementation

Follow docs/salesforce_setup_guide.md for the full build process. In short, the team should:

1. Create the Salesforce Lightning app.
2. Create the four required custom objects and all fields.
3. Configure the Master-Detail and Lookup relationships.
4. Add the three validation rules.
5. Configure the three custom profiles and tab visibility.
6. Set the required organisation-wide defaults.
7. Build the three Salesforce flows.
8. Load the sample records from sample_data/.
9. Capture report screenshots using docs/screenshot_checklist.md.

What To Demonstrate In The Presentation

The demo should focus on the assessed Force.com prototype:

1. The Lightning app and navigation tabs.
2. The four custom objects and their relationships.
3. Submission of a GenAI request by a business user.
4. Validation rules blocking invalid data.
5. Governance review through Risk_Assessment__c.
6. Approval or rejection status updates.
7. Flow-triggered email and task creation.
8. Scheduled deployment reminder design.
9. Optional call to the local FastAPI risk scoring service.

Required Report Screenshots

Use docs/screenshot_checklist.md as the final capture checklist. The report should include screenshots for:

• App home
• All four custom objects
• Relationship fields
• Validation rule errors
• Profiles, object permissions, tab settings, and OWD
• Flow 1, Flow 2, and Flow 3 evidence
• Sample tasks and notifications when available

Recommended Team Workflow

• Build the Salesforce prototype first.
• Use the Python API only after the object model and flows are working.
• Keep screenshots and notes as you configure each Salesforce feature.
• Use the report outline and demo script in docs/ early rather than at the end.