Project detail
AI Data Catalog / Governance Demo
A data catalog product surface for discoverability, ownership, quality, lineage, and approved AI use cases.
Case study
Business Problem and Product Vision
Business Problem
Enterprise AI programs stall when teams cannot quickly tell which datasets are trusted, who owns them, what restrictions apply, and whether a proposed model, copilot, or retrieval workflow is approved. A catalog that only stores metadata does not solve the adoption problem.
Product Vision
This case study turns catalog metadata into a governed product experience: searchable datasets, ownership, sensitivity, quality, lineage, and approved AI-use context in one interface for analytics, compliance, security, and product teams.
Users
Product Users
Live product
Interactive Dashboard
All confidential data is excluded. Portfolio demos use public data or fictional, portfolio-safe sample data to show KPI hierarchy, trust signals, filtering, and regulated-enterprise product workflows.
Fictional AI governance catalog
A product surface for finding trusted datasets, understanding ownership and sensitivity, and confirming which AI uses have been reviewed before a team builds models or retrieval workflows.
Governed domains
9
Business-owned data product areas
Certified datasets
64
Approved for production analytics
Mapped lineage links
97
Source-to-use traceability
Approved AI uses
12
Reviewed model and retrieval patterns
Dataset discovery and AI-use review
| Dataset | Domain | Owner | Sensitivity | Quality | Lineage | Approved AI use case |
|---|---|---|---|---|---|---|
| Customer Golden Record | Customer | Enterprise Data Products | Confidential | 96 | Certified | Personalization and service routing |
| Claims Event Timeline | Insurance Operations | Claims Analytics | Restricted | 91 | Certified | Claims triage assistance |
| Provider Network Access | Healthcare | Network Strategy | Confidential | 88 | Mapped | Access and adequacy analysis |
| Treasury Transaction Signals | Finance and Risk | Risk Data Office | Restricted | 93 | Certified | Anomaly review and investigation |
| Policy Knowledge Base | Legal and Compliance | Governance Office | Internal | 84 | Mapped | Approved retrieval workflow |
| Member Engagement Journey | Customer Experience | Digital Analytics | Confidential | 89 | Mapped | Retention opportunity scoring |
Architecture
Architecture
Reference architecture
Executive view of source, ingestion, analytics, governance, and dashboard delivery.
01
Source systems
Metadata capture
02
Ingestion
Ownership registry
03
Raw / staging
Quality and sensitivity scoring
04
Quality / transform
Lineage graph
05
Product layer
AI-use approval workflow
06
Experience
React / Next.js dashboard
Technical Stack
Data model, pipeline, and implementation
The model treats datasets as governed products with domain, owner, sensitivity, quality, lineage, and approved AI-use attributes. Search and scoring make the catalog useful for daily discovery while retaining the controls required for regulated enterprise AI.
Technology stack
Data Product Decisions
Governance, UX, and architecture choices
What This Demonstrates
What this demonstrates to hiring managers
This project shows how Ted connects governance, metadata, and AI enablement into a practical product interface for healthcare, insurance, banking, and enterprise data teams that need to move faster without losing control of risk.