Use Case: Intelligent Cold Chain Logistics

DEMO

Intelligent Cold Chain Logistics

Experience how AI-driven logistics intelligence improves productivity of FleetOps managers, site reliability engineers and drivers responsible for day-to-day fleet operations delivering pharmaceutical under strict cold chain storage requirements.

Building Blocks used:

ai.agents

automation.optimize

Skills used

Download skills for IBM Bob →

agent
apm-instana

Products Used

1 Review use case

The Smart Cold-Chain Guardrail Platform is an enterprise-grade demo of AI-driven logistics intelligence applied to pharmaceutical fleet operations. It uses a Python simulation engine to model a fleet of 10 refrigerated trucks — each with simulated GPS co-ordinates, temperature telemetry, cargo manifests, and route waypoints — transporting temperature-sensitive goods (vaccines, perishables) against regulatory thresholds (-18 °C frozen / 2–8 °C pharmaceutical). All data is surfaced through an IBM Carbon Design System dashboard with three operator personas: FleetOps Manager, SRE, and Driver.

When a simulated cooling incident is detected — either by the IBM Granite TimeSeries TTM-R2 forecasting service (predicting breach risk up to 96 steps ahead, zero-shot) or by a simulated temperature spike — the platform triggers an autonomous 5-agent pipeline in IBM watsonx Orchestrate: a Weather Advisor scores route-segment severity, a Station Agent finds emergency-cooling facilities within radius, a Route Optimization Agent ranks alternative routes, a Decision Agent produces a risk score (0–100) with financial analysis, and a Notification Agent dispatches a WhatsApp alert via Twilio. The SRE persona monitors infrastructure health via IBM Instana APM and IBM Turbonomic auto-scaling — all deployed on Red Hat OpenShift.

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.

Try the demo

Extra demo assets

Have questions, want a walkthrough or found a bug?

Contact the demo owner for assistance.

Submit a request →

3 Explore how it was built

The demo is built on four interconnected layers, all deployed on Red Hat OpenShift. The Python FastAPI fleetops-backend serves as both the simulation engine and the orchestration hub — it simulates a fleet of refrigerated pharma trucks with GPS coordinates, temperature telemetry, cargo manifests, and route waypoints, performs breach detection, and exposes an Agent Trigger endpoint. A companion IBM Granite TTM-R2 forecast service (forecast-backend) performs zero-shot 96-minute temperature predictions and feeds a WhatsApp/Twilio notification service. When a temperature breach is detected, the FleetOps Manager triggers the agentic workflow from the dashboard UI, which calls the backend's Agent Trigger endpoint; the backend then internally invokes IBM watsonx Orchestrate to run a sequential 5-agent pipeline: a Weather Advisor scores route-segment severity (CLEAR/MOD/SEVERE), a Station Agent locates nearby cold-storage facilities, a Route Optimizer ranks alternative routes, a Decision Agent produces a risk score (0–100) and selects one of four outcomes (EMERGENCY_REROUTE, CONTROLLED_REROUTE, CONTINUE, ABORT), and a Notification Agent dispatches a WhatsApp alert via Twilio. The agent decision is returned to the FleetOps dashboard for the logistics manager to evaluate — the front-end is a Node.js Express application built with IBM Carbon Design System that serves three operator personas — FleetOps Manager, SRE, and Driver — across tabs for Operations, Forecasting, Agents, Observe & Optimize, and Driver. Spanning all layers, IBM Instana collects APM traces and triggers breach alerts, while IBM Turbonomic auto-scales CPU and RAM on latency spikes and persists optimization action history.

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.

Source Code

Jump to: Operations Manager · Driver · SRE

Part 1 — Operations Manager Persona

Demonstrates how IBM FleetOptx ColdPulse uses AI forecasting models and multi-agent systems to detect, analyze, and resolve critical fleet disruptions automatically.

Step 1 — Switch User Persona

Locate the persona dropdown menu in the top-right corner of the interface.
Select Operations Manager – Sarah Chen to access full operational management controls.

Step 2 — Monitor Fleet Operations

Navigate to the main Operations dashboard.
Review the live fleet metrics and active alerts:
- Identify vehicles flagged with critical cooling system threshold breaches.
Click any vehicle displaying a Critical alert in the Alerts sidebar to view its alert history and cargo breach parameters.

Step 3 — Evaluate Predictive Risk (Forecasting Tower)

Click the Forecasting tab in the top navigation bar.
Review the AI-Powered Forecasting summary page:
- View high-risk trucks flagged by the IBM Granite Time Series forecasting model.
- Note the confidence score and projected time-to-spoilage.
Click View Forecast for any truck showing an active breach to analyze detailed predictions:
- Temperature Breach Prediction — historical trends vs. AI future predictions (e.g., 80% risk in next 8 hours).
- Weather Impact Forecast — route conditions and severity levels.
- Station Availability Forecast — available service bays along the route.

Step 4 — Trigger Multi-Agent AI Analysis

Click the blue Analyze button next to an active critical alert.
You will be automatically routed to the Agents tab.
Observe four specialized AI agents executing in parallel:
- Weather Agent — evaluates radar, temperatures, and severe weather delays along route segments.
- Station Agent — searches nearby emergency cooling and repair facilities, available bays, and wait times.
- Route Agent — evaluates alternative rerouting options based on fuel usage, cost, and arrival times.
- Decision Agent — aggregates all findings to generate a unified recommendation, risk score, financial analysis, and post-recovery plan.
Click “Click for full details” on any agent card to drill into its segment data.

Step 5 — Check Autonomous Resolution

In the Decision Agent detail window, review the recommended action: EMERGENCY_REROUTE.
Click Close on the Decision Agent detail window.
Observe the Notification Agent trigger automatically:
- The system reserves an emergency cooling bay at the selected service station.
- An automated WhatsApp notification with updated navigation details is sent to the driver’s device.

Note: For Partner Solution Hub, the notification option is kept disabled during this demo workflow. Please check the video for full functionality.

Part 2 — Driver Persona

Shows how IBM FleetOptx ColdPulse supports the Driver Persona — automatically processing critical telemetry alerts, rerouting drivers to emergency cooling hubs, and presenting clear actionable instructions without manual intervention.

Step 1 — Switch User Persona

Locate the persona dropdown menu in the top-right corner of the interface.
Select Driver Persona to open the dedicated Driver View portal.

Step 2 — Select Driver Profile

Locate the Select Driver Profile dropdown in the top-left section of the dashboard.
Select the driver/truck profile displaying a Critical alert status to view its active telemetry.

Step 3 — Review Vehicle & Cargo Telemetry

Navigate to the My Truck section in the central panel.
Monitor real-time status:
- Cargo Temp — observe active temperature alerts (highlighted warning levels).
- Vehicle Status — check fuel levels, route progress, remaining distance, ETA, speed, and overall cargo status.

Step 4 — Access Reroute Notifications

Locate the Notifications panel on the right side of the screen.
Review the automated EMERGENCY_REROUTE notification:
- Updated destination details (e.g., assigned emergency service plaza).
- Facility proximity, estimated arrival window, cargo temperature recovery buffer, weather clearance, and available docking status.

Step 5 — Verify Rerouted Destination & Navigation

Check the Destination Station panel to confirm assigned facility details:
- Route & Station Type — station location, service type, and coordinates.
- Status & Bay Availability — confirm an active service bay is reserved for emergency cooling.
Reference the My Current Route map panel on the left to trace the live GPS trajectory updated with the new service plaza destination.

Step 6 — Review Delivery History

Scroll to the My Recent Deliveries panel at the bottom of the interface.
Review past completed routes, cargo categories, total distances, and duration logs to verify historical operational compliance.

Part 3 — SRE Persona

Note: For Partner Solution Hub, application observability is kept disabled during this demo workflow. Please check the video for full functionality.