SourceGuard

** Please Check the Try out link for Demo Video**

Inspiration

The "Greenwashing" crisis. Corporations are under immense pressure to meet ESG goals, often paying a "Green Premium" for sustainable materials. However, supply chain fraud is rampant and invisible because data is siloed: Finance sees the invoice, but Operations sees the waste. We wanted to bridge this "Trust Gap" to protect both the bottom line and the planet.

What it does

SourceGuard is a Supply Chain Command Center. It correlates Procurement Spend (Finance) with Physical Carbon Intensity (Logistics) to detect "Value Leakage."

• Visualizes Risk: A "Greenwashing Matrix" instantly flags vendors who charge high prices but deliver high-carbon, inefficient shipping.
• Calculates Efficiency: Uses a custom "Carbon Efficiency Ratio" to quantify environmental performance per dollar spent.
• Automates Action: Users can ask Agentforce to analyze the data and draft legal "Notice of Breach" letters or remediation plans instantly.

How we built it

We leveraged the full Tableau Next stack to create a cognitive application:

1. Data Cloud: Ingested disparate datasets (Financial Contracts and Logistics Logs).
2. Semantic Layer: Built a multi-object model to link VendorID across silos and defined the "Carbon Efficiency Ratio" metric with AI-readable descriptions.
3. Visuals: Designed a "Kill Zone" scatter plot using reference lines to isolate non-compliant vendors.
4. Extensibility: Used the App Template Framework to package the dashboard as a deployable JSON template for the Marketplace.

Challenges we ran into

Aligning data granularity was difficult. Financial data exists at the "Contract" level, while Logistics data exists at the "Shipment" level. We utilized the Semantic Model's relationship capabilities to normalize these values, ensuring the calculated metrics were accurate across different shipment modes (Air vs. Ocean).

Accomplishments that we're proud of

We didn't just build a dashboard; we built a Decision Engine. We are most proud of training Agentforce to interpret our custom metrics. Seeing the Agent correctly identify a fraudulent vendor and draft a termination letter based solely on our Semantic Model definitions was a huge win.

What we learned

Context is everything. We learned that in the age of AI, the Semantic Layer is the most critical component. Without clear synonyms and descriptions, the AI is blind. With them, it becomes an expert analyst.

What's next for SourceGuard

• ERP Integration: Automating "Payment Blocks" in Salesforce when a vendor hits the high-risk zone.
• Scope 3 Expansion: Adding labor compliance data to create a holistic "Ethical Sourcing" score.

Built With

• datacloud
• einstein
• tableaunext