TL;DR
A shipment does not wait for a database to update before it moves.
A truck leaves a supplier, a container reaches a port, goods clear customs, inventory enters a warehouse, and an order reaches the customer. Each step generates data, but that data often moves slower than the physical supply chain itself.
Many organisations still rely on EDI, scheduled integrations, file transfers, and periodic database synchronisation. Each system may have an accurate view of its own operations, but there is often no continuously updated view across the entire supply chain.
That creates blind spots.
A supplier delay may only become visible when a delivery is already late. A warehouse may discover an inventory shortage after demand has increased. A customer may still see an order as "in transit" even after a delivery exception has occurred.
The problem is not a lack of data. It is the time between an event happening and the business being able to act on it.
Kafka addresses this by turning supply chain changes into events that can flow continuously between systems. Shipment updates, inventory changes, supplier events, warehouse transactions, and IoT data can be processed as they happen.
But streaming the data is only the starting point. The real value comes from applications that can interpret those events, identify risks, and trigger actions.
This is where Condense Kafka and Condense Applications fit together. Condense Kafka provides the managed streaming foundation, while Condense Applications provides the environment to build and operate the real-time processing and business logic on top of it. Its BYOC model also allows the platform to run within the customer's cloud environment.
The result is a shift from simply knowing what happened to understanding what is happening now and what needs to happen next.
The Problem: EDI and Batch Supply Chain Data Creates Blind Spots
Supply chains operate continuously, but enterprise data often moves in batches.
A supplier can change a production schedule, a truck can be delayed, or warehouse inventory can fall below its threshold, while downstream systems continue working with an older view of the situation.
This creates a gap between what is happening physically and what the business knows digitally.
The 24-Hour Visibility Gap
Consider a shipment expected at a distribution centre on Tuesday. The supplier sends an EDI confirmation and the warehouse plans around the expected arrival. But the truck is delayed.
If the next update reaches the warehouse only during the next scheduled integration cycle, the team may continue planning against an arrival that is no longer realistic.
By the time the delay becomes visible, it may already affect inventory, production, or customer commitments.
Why Batch Integration Falls Short
Batch processing works well when data does not need an immediate response. Supply chain events often do.
A missed production milestone, route deviation, inventory shortage, temperature breach, or failed delivery attempt can require action within minutes, not at the end of the day.
Point-to-point integrations also become harder to manage as more systems are added. ERP, WMS, TMS, supplier platforms, and customer applications can end up with multiple independent connections, increasing both complexity and maintenance.
Moving to Event Streaming
Kafka provides a different model. Instead of periodically exchanging the latest state, systems can publish meaningful changes as events to a shared streaming layer.
A shipment dispatch event, for example, can simultaneously update the TMS, WMS, ERP, analytics applications, and customer portal.
This is where Condense Kafka can provide the managed streaming foundation. With BYOC, the streaming platform can run within the customer's cloud while providing the operational benefits of managed Kafka.
The result is a simple but important shift:
Instead of waiting for the next data update, supply chain systems can respond to events as they happen.
Event Types in a Real-Time Supply Chain
A supply chain is made up of thousands of events that continuously change the state of orders, inventory, shipments, warehouses, and deliveries.
For real-time visibility, the goal is not to stream everything. It is to identify the events that matter and make them available to the systems that need to act on them.
Purchase Order Events
The process starts with procurement events such as:
Purchase order created or approved
Supplier acknowledgement
Quantity or specification changes
Delivery date changes
Purchase order cancellation
These events establish what is expected to happen next.
Supplier and Production Events
Supplier systems add events such as production started, production completed, material shortages, quality checks, shipment readiness, and delays.
On their own, these events provide status. When combined with inventory and demand data, they can indicate potential supply risk.
Shipment and Transportation Events
Transportation generates events throughout the journey:
Shipment dispatched
Vehicle assigned or departed
Route deviation
Geofence entry or exit
Port or hub arrival
Customs clearance
Delivery exception
Shipment delivered
IoT and telematics can enrich these events with location, temperature, vehicle status, and other operational data.
Warehouse and Delivery Events
Once goods reach the warehouse, events include goods receipt, inventory movement, picking, packing, and dispatch.
The final mile adds events such as out for delivery, delivery delay, failed attempt, rescheduling, proof of delivery, and completion.
The real value comes from connecting these events.
A supplier delay combined with low inventory, increasing demand, and a delayed shipment tells a very different story from any one of those events alone. A real-time application can correlate them and generate a higher-level event such as supply risk detected.
This is where the streaming layer moves beyond data transport. With Condense Kafka providing the managed event backbone, Condense Applications can help to build business logic to process and correlate events to turn operational data into supply chain intelligence and actions.
Architecture: IoT Sensors → MQTT → Kafka → Supply Chain Visibility Layer
Real-time supply chain visibility becomes more useful when it includes data from the physical world.
Vehicles, containers, cold-chain equipment, warehouses, and other assets can continuously generate events such as location, temperature, humidity, vehicle status, and asset movement.
For these environments, MQTT and Kafka complement each other well. MQTT provides lightweight communication from devices and gateways, while Kafka provides the scalable event streaming layer for processing and distributing those events across enterprise applications.
From Telemetry to Business Events
Raw IoT data is not always meaningful to a supply chain application.
A GPS update may simply contain coordinates. A streaming application can turn that into:
Vehicle entered distribution centre
Similarly:
Temperature above threshold
can become:
Cold-chain exception detected
This transformation is important because downstream systems typically need business context rather than raw telemetry.
Why Kafka Sits Between Devices and Applications
Sending device data directly to every application creates multiple integrations and makes the architecture harder to scale.
With Kafka, an event can be published once and consumed by multiple applications, such as a fleet management system, TMS, warehouse application, analytics platform, or customer portal.
This also makes it easier to introduce new applications without changing the underlying device integration.
Where Condense Fits
Condense Kafka can provide the managed streaming layer for IoT, MQTT, and enterprise events, while Condense Applications can help to build the business logic to process that data using business rules, transformations, and custom application logic.
For example, vehicle telemetry can be enriched with shipment information, converted into a delivery event, and then consumed by operations or customer-facing applications.
Condense's platform combines data ingestion, processing, integrations, visualisation, and real-time application capabilities, with BYOC allowing the platform to operate within the customer's cloud environment.
This creates a practical path from device data to supply chain intelligence, without making every downstream application responsible for understanding raw IoT data.
Demand Sensing: Correlating POS Events With Upstream Inventory in Real Time
Demand can change much faster than traditional supply chain systems can respond.
A product that normally sells 100 units an hour can suddenly see demand double, while warehouse inventory falls and the next supplier shipment is delayed. Each system may show a different part of the picture, but the risk becomes clear only when these events are correlated.
Kafka provides the common event stream to bring these signals together. POS and eCommerce events can be combined with inventory updates, supplier events, and shipment status to continuously calculate demand versus available supply.
For example, if inventory is 1,200 units, demand is 200 units per hour, and the next replenishment is expected in 10 hours, the business has only six hours of inventory coverage. That is an early warning for a potential stockout.
This is where Condense Applications adds value on top of Condense Kafka. A real-time application can be built using the ai assistance or any code of choice or connect the git repositories to build and deploy applications which calculate demand velocity, track inventory coverage, consider replenishment timelines, and generate an alert when risk thresholds are crossed.
The response can then be operational:
Prioritise replenishment
Alert procurement
Redistribute inventory
Adjust production planning
Escalate a potential stockout
The shift is from asking "How much inventory do we have?" to "How long will our inventory last at the current rate of demand?"
With streaming data, that answer can continuously change as new orders, inventory movements, supplier updates, and shipment events arrive.
Supplier Risk Monitoring: Detecting Delays Before They Become Stockouts
A supplier delay is rarely just a supplier problem. If the affected material is critical, it can quickly impact production, warehouse inventory, transportation, and customer commitments.
The challenge is identifying that risk early enough to act.
A real-time streaming architecture can continuously correlate supplier updates with shipment status, inventory levels, demand, and expected delivery times. For example, a delayed production milestone may become significantly more important when the related shipment has not been dispatched and warehouse inventory is already approaching its safety threshold.
Instead of treating these as separate updates, a streaming application can turn them into a single business signal:
Supplier delay + low inventory + high demand = supply risk
That signal can trigger actions such as:
Alert procurement
Escalate to operations
Request an updated supplier ETA
Prioritise an alternate supplier
Reallocate inventory
Adjust production planning
This is where Condense Applications can work on top of Condense Kafka. Kafka provides the continuous event stream, while the application layer can implement the organisation's supplier-risk rules and evaluate them as new events arrive.
The advantage is context.
A supplier with a minor delay may not require intervention when inventory is healthy. The same delay can become critical when only a few hours of inventory remain.
Real-time streaming therefore moves supplier monitoring from "Is the supplier late?" to "Will this delay impact our operations, and what should we do now?"
That is the difference between simply monitoring supplier events and detecting supply chain risk early.
Last-Mile Visibility: Streaming Delivery Events to Customer-Facing Applications
The last mile often has the biggest impact on customer experience. A shipment may move smoothly through suppliers, warehouses, and transportation hubs, only for a delivery delay or failed attempt to create a customer escalation.
Real-time streaming can connect these delivery events directly to customer-facing applications.
Events such as out for delivery, delivery delayed, vehicle approaching, failed delivery attempt, rescheduled, and delivered can flow through Kafka as they occur. Instead of a customer portal repeatedly polling the TMS or order management system, it can consume the relevant events and maintain an up-to-date shipment status.
The same event can also serve multiple applications. A delivery exception can update the customer portal, alert operations, update the TMS, and trigger a customer notification.
The key is adding context to raw delivery data.
A vehicle entering a delivery zone can become "Your order is arriving soon." A prolonged unexpected stop can become "Your delivery may be delayed."
Condense Kafka provides the streaming foundation for these events, while Condense Applications can implement the logic that connects telemetry, shipment information, delivery windows, and customer status.
This creates a more responsive last-mile experience without requiring every application to build its own connection to the underlying transportation systems.
Integration Patterns: Kafka → ERP, WMS, TMS, Customer Portal
Kafka does not need to replace the systems already running the supply chain. Its role is to provide a real-time event layer between them.
Kafka → ERP
ERP systems can publish and consume events such as purchase orders, goods receipts, inventory updates, and shipment completion. This keeps the ERP as the system of record while making important changes available to other applications in real time.
Kafka → WMS
Warehouse events such as goods receipt, inventory movement, picking, packing, and dispatch can be streamed to other systems. An inventory threshold event, for example, can trigger a replenishment workflow without waiting for a scheduled update.
Kafka → TMS
Transportation events such as vehicle assignment, dispatch, ETA changes, delivery exceptions, and proof of delivery can be streamed to operations, analytics, and customer applications. Applications can also publish events back to the TMS when a business condition requires action.
Kafka → Customer Portal
Customer applications typically need only selected events. A streaming application can consume relevant shipment events and provide a continuously updated status such as dispatched, in transit, delayed, out for delivery, or delivered.
One Event, Multiple Uses
The same event can serve several downstream systems without creating another point-to-point integration. This makes it easier to add new supply chain applications while keeping the existing ERP, WMS, and TMS systems in place.
With Condense Kafka, organisations can use managed Kafka as this shared event backbone, while Condense Applications can handle the transformation, enrichment, and business logic between streaming events and enterprise workflows. Condense's platform includes integrations, processing, and application capabilities designed for these real-time workloads.
Where Condense Fits: Managed Kafka and Real-Time Supply Chain Applications
Kafka provides the event backbone, but running it reliably at production scale still requires infrastructure, monitoring, scaling, upgrades, security, and ongoing operations.
This is where Condense Kafka fits. It provides a managed Kafka foundation for continuous supply chain events while supporting deployment within the customer's cloud through BYOC. This gives enterprises the operational simplicity of managed Kafka without moving their streaming environment outside their cloud.
But the streaming backbone is only one part of the architecture.
Condense Applications provides the layer for building real-time business logic on top of those events. Teams can develop applications for supplier risk, demand sensing, shipment tracking, inventory alerts, ETA prediction, or delivery exceptions without treating each use case as a separate infrastructure project.
The distinction is simple:
Condense Kafka manages the event streaming foundation.
Condense Applications turns those events into business logic and actions.
This also allows organisations to build on the same streaming foundation over time. A shipment event used today for tracking could later power an ETA application, supplier-risk model, customer notification workflow, or anomaly detection application.
The result is a more practical path to modernising supply chain data. Existing ERP, WMS, and TMS systems can continue doing what they already do, while Condense provides the real-time streaming and application layer connecting them.
The goal is not to add another system to the supply chain. It is to create the real-time layer that allows the systems already there to work together more intelligently.
From Supply Chain Visibility to Real-Time Supply Chain Intelligence
The real value of real-time supply chain streaming is not simply knowing where a shipment is.
It is knowing what is happening across the supply chain, understanding why it matters, and giving the right system enough time to respond.
That requires connecting events that traditionally live in separate systems.
A supplier delay can be correlated with inventory levels. Inventory can be correlated with current demand. Demand can be correlated with shipment ETAs. Shipment events can be correlated with customer commitments.
The result is a much richer operational picture.
From "What Happened?" to "What Is Happening?"
Traditional reporting is often retrospective.
A report might tell the supply chain team:
12% of shipments were delayed last month.
Useful, but the information arrives after the opportunity to intervene has passed.
A streaming architecture can provide a different question:
Which shipments are becoming at risk right now?
That change is important.
Instead of analysing yesterday's supply chain, teams can continuously evaluate the current state of the operation.
From Visibility to Action
Once events are available in real time, the next step is to turn them into actions.
For example:
Supplier delay detected
→ Recalculate expected inventory
Inventory coverage falls below threshold
→ Alert procurement
Demand increases unexpectedly
→ Recalculate replenishment requirement
Shipment ETA exceeds customer commitment
→ Trigger delivery exception
Vehicle reaches delivery zone
→ Update customer application
Each action can be driven by an event rather than a scheduled job.
This is where the combination of Condense Kafka and Condense Applications becomes particularly relevant.
Condense Kafka provides the managed event streaming foundation, while Condense Applications can host the custom processing, business rules, and real-time workflows built around those events. The platform's application runtime supports the lifecycle from development and validation through deployment, orchestration, monitoring, and operation.
A Foundation for AI-Ready Supply Chains
Once supply chain events are continuously available, the same streaming foundation can also support more advanced intelligence.
Historical events can provide context for models, while live events can provide the latest operating state.
This opens the door to applications such as:
Predictive ETA
Demand forecasting
Supplier risk scoring
Inventory optimisation
Route anomaly detection
Delivery prediction
Equipment anomaly detection
The important point is that AI does not have to sit separately from the operational data flow.
Models can become part of the streaming application layer, consuming events and producing new events or recommendations as conditions change.
Condense's platform material positions the platform as AI/ML ready, with real-time application development and processing capabilities designed to operate alongside streaming data.
The End State: A Supply Chain That Responds in Real Time
The progression is straightforward:
That is the real opportunity behind real-time supply chain visibility with Kafka.
Kafka provides the event backbone. A platform such as Condense makes that backbone easier to operate and extends it into the application layer where supply chain-specific logic can be built and run.
The result is not simply a faster data pipeline.
It is a supply chain that can see, understand, and respond while events are still unfolding.
Conclusion: Building a Truly Real-Time Supply Chain
Supply chains do not operate in batches.
Suppliers produce continuously. Vehicles move continuously. Warehouses receive and dispatch continuously. Customers place orders continuously.
The data describing these activities should not have to wait for the next scheduled sync.
Kafka provides the foundation for changing that model. By turning supply chain activities into events and making those events continuously available, organisations can connect suppliers, transportation, warehouses, ERP systems, customer applications, and IoT devices through a common streaming layer.
But real-time visibility is only the beginning.
The real opportunity comes from building applications that can interpret those events and act on them. A supplier delay can become a risk signal. A change in demand can become a replenishment trigger. A vehicle event can become a customer notification. An inventory change can immediately influence planning.
This is where Condense Kafka and Condense Applications fit together.
Condense Kafka provides the managed streaming backbone, while Condense Applications provides the environment to build and operate the real-time processing and business logic that sits on top of those streams. With BYOC, the platform can also operate within the customer's cloud environment, supporting organisations that need control over their data and infrastructure.
The result is a shift from:
Batch → Streaming
Visibility → Intelligence
Alerts → Action
Reactive supply chain → Responsive supply chain
A modern supply chain should not discover a disruption after it has already affected production or customers.
It should see the signal early, understand the context, and have the systems in place to respond.
That is the real promise of real-time supply chain visibility with Kafka.
