supplygraph-ai

Procurement & Inventory Risk – Key component Focus

Overview

The Procurement & Inventory Risk – Key component Focus continuously monitors the MPNs you cover. It watches whether supply-chain events and market signals are moving price, lead time, and market inventory for that specific part.

Fragmented events, path evidence, and distributor quotes are turned into structured, auditable, and continuously updated MPN-level risk warnings, so you can see which components are under pressure and keep tracking how that pressure develops.

Pain

Procurement and supply-chain teams need 24/7 visibility into critical MPNs — not only that a global event occurred, but whether this part is being hit, on which metric, and by how much.

Without MPN-level monitoring, teams are often left with:

• Event alerts that never resolve to the parts they actually buy
• No clear account of why an upstream shock may transmit to a given MPN
• A generic warning instead of quantified movement in price, lead time, and market inventory
• A first hit that is not followed as the situation develops

Traditional monitoring can surface an event. It usually cannot say whether that event is already showing up in this MPN’s market, or keep following the move after the first notice.

Breakthrough

The Supply Chain Risk Prediction – Cross-Product Agent extends supply chain monitoring into product-driven risk propagation.

Behind the scenes, it:

• Monitors global events affecting upstream products or services
• Quantifies how an original event impacts a specific product or service
• Identifies product or service dependency paths between upstream inputs and the target company
• Propagates the product-level affected state toward the target company
• Estimates downstream enterprise risk, confidence, arrival timing, and key transmission logic
• Returns structured warning events and actionable recommendations

For the user, the workflow remains simple:

1. Provide a target company and an upstream product or service
2. Confirm the matched entities
3. Activate monitoring for the product under the target company
4. Retrieve the latest product-driven risk warnings at any time

Typical insights include:

• Product-level event impact
• Risk transmission path
• Enterprise risk value and change
• Confidence and confidence change
• Estimated arrival days
• Key price or market signals
• Actionable mitigation recommendations

Why SupplyGraph AI

Powered by:

• 100M+ global companies
• 1M+ key products and components
• Real-time global signal capture running 24/7
• Enterprise-product dependency graph intelligence
• Deep-tier product-level risk quantification engine

SupplyGraph AI provides a continuously updated, auditable product-driven supply chain risk intelligence layer that connects upstream product disruptions to downstream enterprise exposure.

This makes the agent suitable for:

• Supply chain risk teams
• Procurement and sourcing teams
• Enterprise risk management teams
• Investors and analysts
• Global operating companies with critical upstream product dependencies

Agent Behavior Model

Sandbox Key Support (for Development)

This agent supports Sandbox API Keys, enabling developers to test integrations without consuming credits.

When using a Sandbox Key:

• No credits are deducted
• The agent returns predefined sample product-driven risk warning data
• Data is static and does not update over time
• Responses do not trigger live monitoring or live risk computation
• Only validation and schema checks are executed

Sandbox Keys are ideal for:

• UI prototyping
• API integration testing
• SDK development
• CI/CD pipelines

⚠️ Sandbox data must not be used for production analytics.

The full A2A workflow (run → status → result) behaves the same as production, enabling accurate integration testing.

Initial Monitoring Setup

When a user initiates a new monitoring request through /run without specifying mode, or with mode=run, the agent starts the product-level monitoring customization process for a specific target company.

The request should provide both:

• Target Company
• Upstream Product or Service

The agent then identifies and returns the matched entities for user confirmation.

First-time workflow:

1. User calls /run with target company and upstream product or service information
2. System creates a task_id
3. Agent returns matched entities for confirmation
4. User confirms the matched entities by replying Yes
5. The product monitoring customization is activated
6. The one-time customization fee is charged
7. User retrieves the latest warning data via mode=result

After activation:

• The agent runs continuously
• Results are updated in real time
• Each result request returns the latest available warning data
• Status becomes completed after the initial setup is finished

Pricing & Billing Model

Pricing is based on product-level customization under a target company.

• Adding a Target Company is free
• Customizing an upstream product or service under a target company is paid
• The current price is $4,999 per upstream product per target company
• This is a one-time fee, not a subscription

The fee is charged only after the user confirms the matched target company and upstream product or service.

The fee covers:

• Product-level monitoring under the specified target company
• Continuous event detection
• Product-driven risk propagation analysis
• Latest warning data retrieval through the result mode
• Structured and auditable warning outputs

Understanding task_id and Monitoring Lifecycle

A task_id represents a product monitoring instance under a specific target company.

How task_id is created and maintained

• /run without task_id creates a new task
• After entity confirmation, the task becomes an active monitoring instance
• The task remains valid for retrieving the latest warning data
• The system recognizes the monitoring instance by task_id

Implications:

• If a user loses the task_id and initiates a new /run, the system treats it as a new customization request
• If the same product needs to be monitored under multiple target companies, each target-product pair is treated independently
• Each upstream product is priced independently under each target company

Status Endpoint Behavior

The status mode is used to check whether the task setup has completed.

For this agent:

• Once the initial setup is completed, the task status remains completed
• Ongoing risk updates should be retrieved through mode=result
• Repeated status polling is usually unnecessary after completion

Result Endpoint Behavior

The result mode always returns the latest available warning data for the activated target-product monitoring instance.

• It may be called repeatedly
• Each request returns the freshest available data
• The agent continues running and updating warning events in the background
• The response may contain one or more active warning events

Input Requirements

agent_id: supply_chain_risk_prediction_cross_product · MCP tool: supply_chain_risk_prediction_cross_product

Field	Required	Description
text	Yes	Target company and supplier/component information

text Format

The text field should include both target company and upstream product or service information.

Recommended format:

target company: company name, country/region, address(optional), website(optional); upstream product: product name (or MPN-XXXXX), company name, country/region, address(optional), website(optional)

Example:

target company: Tesla, Inc. United States; upstream product: Lithium Carbonate Albemarle Corporation United States

Example Response (WAITING_USER)

{
  "success": false,
  "code": "WAITING_USER",
  "data": {
    "task_id": "<task-id>",
    "agent": "supply_chain_risk_prediction_cross_product",
    "stage": "interpreting",
    "content": "Here are the entities we’ve identified.\nTarget Company Name: Tesla, Inc.\nCountry: United States\nUpstream Product: Lithium Carbonate\nCompany Name: Albemarle Corporation\nCountry: United States\nPlease reply [Yes] or [No] to confirm."
  }
}

After the user replies Yes, the product monitoring customization is activated and the one-time fee is charged.

Confirm matched entities when prompted (WAITING_USER). Uses mode=result for latest data.

Integration

Method	ID / Tool	Documentation
A2A	supply_chain_risk_prediction_cross_product	a2a.md
MCP	supply_chain_risk_prediction_cross_product	mcp.md
Agent API	supply_chain_risk_prediction_cross_product	agent-api.md

Quick examples: A2A / MCP · Agent API

Errors

Common codes → Agent API §10. Returns WAITING_USER for company confirmation. Uses mode=result.

Maintainer: info@supplygraph.ai License: Proprietary / Internal © 2025–2026 SupplyGraph AI. All rights reserved.

This site is open source. Improve this page.