DEMO

Real-Time Supply Chain Risk Control Tower with Confluent

Correlates live inventory, shipment, and external-risk signals to detect supply-chain disruption early, score business risk continuously, and recommend mitigation before customers are impacted.

Building Blocks used:

ai.agents

data.pipelines

Skills used

Download skills for IBM Bob →

Products Used

1 Review use case

Supply-chain risk rarely exists in one application. Inventory, supplier, logistics, and external-risk signals are usually fragmented across systems, so teams discover issues through batch reports, spreadsheets, and email after the disruption has already grown. In the demo, NorthStar Energy Systems has a strategic 900-unit order worth $750K with a four-day ship window and a critical part, BRG-9004, at risk.

IBM Bob orchestrates a reusable Confluent Building Block that streams inventory, shipment, and port-risk events, while a risk engine correlates the signals and continuously recalculates the business risk. The scenario progresses from MEDIUM to HIGH to CRITICAL and then back to MEDIUM as conditions recover. When risk becomes critical, the solution publishes risk and alert events, recommends actions such as activating an alternate supplier and expediting the shipment, and uses IBM watsonx.ai to draft an executive summary and supplier escalation communication for human approval.

Start Here

Watch this short overview

Preview the application

Explore key screens from the demo

2 Try the interactive application

Experience the interactive demo and explore its features. A demo script is provided to help when presenting the application to partners.

Have questions, want a walkthrough or found a bug?

Contact the demo owner for assistance.

Submit a request →

3 Explore how it was built

Source systems such as ERP/OMS, supplier portals, shipment tracking, inventory systems, and external-risk feeds publish business events through a reusable Confluent Building Block using connectors/producers, Kafka topics, and Schema Registry. IBM Bob orchestrates the demo flow from a spec-driven task plan. A Python or Flink SQL risk engine joins the signals, calculates days of supply and a multidimensional risk score, and publishes risk-scores, recommendations, and alerts; IBM watsonx.ai generates the explanation and communications, which can then drive dashboards, alerts, procurement workflows, or analytics.

Architecture

4 Download and modify with IBM Bob

Download the source code and specifications for this application to set up and run it locally. Once downloaded, open the code in IBM Bob to modify and adapt it to your own use case.

Specifications

Source Code

Real-Time Supply Chain Risk Control Tower with Bob and Confluent — Demo Script

Overview

NorthStar Energy Systems has a strategic 900-unit order worth $750K with a four-day ship window. A critical part, BRG-9004, is at risk. This demo walks through how IBM Bob, Confluent, and IBM watsonx.ai correlate live signals, score business risk continuously, and recommend mitigation before customers are impacted.

Step 1 — Set the Scene with the Strategic Order

Present the customer order context:

Show the 900-unit order totalling $750K revenue with a four-day ship window.
Highlight the critical part BRG-9004 and explain why it is the single point of failure.
Explain that without signal correlation, the risk is invisible in any single system.

Step 2 — Show the Three Independent Risk Signals

Walk through each signal in isolation:

Show low inventory at the plant — insufficient units to fulfil the order without resupply.
Show the supplier shipment delay increasing from 12 hours to 96 hours — a 8× degradation.
Show severe port congestion on the primary logistics route — blocking the fastest path to the customer.

Emphasise that each signal alone looks manageable, but their combination is what drives the escalation.

Step 3 — Show the Confluent Streaming Backbone

Demonstrate real-time event correlation:

Open the Confluent topics view and show inventory, shipment, and port-risk events streaming in real time.
Explain that the risk engine subscribes to all three topics simultaneously and recalculates a multidimensional risk score on every new event.
Point out that this is the reusable Confluent Building Block orchestrated by IBM Bob — new signal sources can be added without redesigning the pipeline.

Step 4 — Watch the Risk Escalate

Trace the risk lifecycle from MEDIUM to CRITICAL:

Show the business risk indicator move from MEDIUM → HIGH as the shipment delay worsens.
Show it escalate to CRITICAL as port congestion combines with the inventory shortfall.
Highlight the continuous recalculation — this is not a batch report; risk is visible the moment conditions change.

Step 5 — Review the Recommended Mitigation at CRITICAL

Show the automated recommendation engine:

At CRITICAL risk, show the three recommended actions published to the alerts topic:
- Activate alternate supplier SUP-221 to bypass the delayed primary supplier.
- Expedite shipment SHP-33019 to recover the delivery timeline.
- Notify procurement to initiate purchase order approval for the alternate source.
Explain that these recommendations are derived from the correlated risk score, not hardcoded rules.

Step 6 — Show IBM watsonx.ai-Generated Communications

Demonstrate AI-assisted human communication:

Open the Executive Summary draft generated by IBM watsonx.ai — a concise briefing of the risk situation, financial exposure, and recommended actions.
Open the Supplier Escalation Communication — a pre-drafted message to the alternate supplier SUP-221 ready for human approval and send.
Explain that a human approves before anything is dispatched — AI drafts, humans decide.

Step 7 — Replay the Recovery Scenario

Show the system returning to normal:

Trigger the recovery conditions: shipment delay returns to 12 hours and inventory rises to 820 units.
Watch the risk indicator automatically return to MEDIUM as the risk engine recalculates.
Emphasise that the recovery is as automatic and continuous as the escalation.

Step 8 — Close with the Reusable Pattern

Highlight extensibility:

Explain that the pattern is specification-driven: adding a new signal (weather, tariffs, supplier quality scores) requires only:
- Updating the spec.
- Adding an event contract to the Schema Registry.
- Replaying the acceptance scenario in IBM Bob.
Show the architecture diagram and point out the clean separation between the streaming backbone, the risk engine, and the IBM watsonx.ai communication layer.
Summarise: any industry with multi-source risk signals — retail, pharma, defence — can adopt this pattern.