Dategro logo

Global Trust Challenge · Phase 2 · dategro IT

Trust, Compliance, and AI: A Unified Approach

A sales order leaves an ERP system, travels through a transport layer and is processed by an AI forecast. Can you still trust the result? This prototype shows each dataset's origin, integrity, processing history and governance status, recorded as signed trust events along the way and evaluated in one dashboard.

The system does not judge whether a forecast is correct. It proves where the data came from and whether anyone tampered with it on the way.

Start the demo Open Trust Dashboard
  • Origin
  • Integrity
  • Processing history
  • Governance status

The data lifecycle

One dataset, five stations, one verdict

Every station that touches the dataset writes a signed trust event to the Trust API. The three highlighted stations are live apps in this demo, so click one to open it.

  1. ERP Order ERP Simulator ENTER Open ERP →
  2. IDLS Transfer Transport layer TRANSFER
  3. AI Forecast AI system TRANSFER · EXIT Open AI-Forecast →
  4. Forecast Result Dataset ERROR if manipulated
  5. Auditor Review Trust Dashboard EVALUATE Open Dashboard →

How to use the demo

Three apps, three steps

Follow the steps in order. Each app runs on its own and uses predefined, reproducible data, so you can repeat a run at any time.

1

ERP Simulator

The data source. It is a lightweight ERP in an SAP UI5 / Dynamics look that creates the sales orders everything else depends on.

How to use it

  1. Pick one of the five predefined sales orders, e.g. SO-2026-2001 Standard domestic or SO-2026-2005 Nordic cross-border.
  2. Review the header, line items and process flow.
  3. Select it, tick Send as signed source and click Send to Server to hand the order over to the trust chain.

What you see

  • The system identity card of the issuing ERP (svc-erp-simulator).
  • The trust event panel with the signed ENTER event that marks the dataset's origin.
Open ERP Simulator →
2

AI Forecast

The processing step. An AI system reads the submitted orders and produces a forecast. This is where the data can be manipulated.

How to use it

  1. Select the order group you just submitted. The order data is resolved live from the ERP.
  2. Choose a path:
    • Create Forecast processes the data unchanged.
    • Create invalid Forecast changes an amount on the way, which simulates tampering.

What you see

  • The resolved orders and the forecast details.
  • The TRANSFER / EXIT events that are written. On the manipulated path an ERROR event breaks the chain.
Open AI Forecast →
3

Trust Dashboard

The auditor's view. It collects every trust event for a dataset and turns them into one clear, explained verdict.

How to use it

  1. Pick the dataset from the selector at the top.
  2. Read the overview, then drill into the evidence.

What you see

  • The overall trust status and the lifecycle strip.
  • Five trust dimensions: Source Identity Cryptographic Integrity Chain Completeness Processing Authorisation Governance Compliance
  • A plain-language explanation and the raw evidence table of all events.
Open Trust Dashboard →

Four Scenarios

Same order, different outcome

Complete Chain

✓ Verified

Submit an order in the ERP, create a valid forecast and open the dashboard.

ENTER→ TRANSFER→ EXIT

The chain is complete and every signature checks out. All five dimensions pass.

Manipulation

✕ Compromised

Submit the same order, but with a man-in-the-middle attack.

ENTER→ TRANSFER→ ERROR

The data no longer matches what the ERP signed, e.g. due to an attack. The dashboard flags the break and explains which dimension failed.

Incomplete Provenance

✓ Partially Verified

Submit the same as an unsigned order.

ENTER→ TRANSFER→ EXIT

The order comes from a source that cannot sign and receives a gateway attestation

Manipulation by System

✕ Compromised

Submit the same order, but create an invalid forecast that changes an amount.

ENTER→ TRANSFER→ ERROR

The data no longer matches what the ERP signed. The dashboard flags the break and explains which dimension failed.

Status legend Verified Partially Verified Compromised Unknown

Ready? Start where the data starts.

Create an order in the ERP, run it through the AI and watch the dashboard reach its verdict.

1 · ERP Simulator 2 · AI Forecast 3 · Trust Dashboard

Behind the transport layer

Trust infrastructure

Four building blocks make every trust event verifiable, from the identity of the sender to the final status in the dashboard.

  1. Trust Registry & Key Management The Trust Registry maintains the identities of participating systems and the public information required to verify their signatures.
  2. Signing & Verification Service The Signing and Verification Service provides cryptographic functions for the prototype components.
  3. Trust Evidence Store The Trust Evidence Store persists the trust events generated throughout the data lifecycle.
  4. Evaluation Engine The Evaluation Engine derives a rule-based trust status from the stored trust events.