The Problem
EpicFlow’s clients — logistics and industrial operations teams — faced the same problem from opposite directions. Operations staff who understood the business processes they needed to automate could not implement automations without developer support. Developers who could implement automations didn’t understand the domain well enough to design them correctly without extensive specification work.
The gap between domain knowledge and technical execution was creating backlogs, misimplemented automations, and processes that worked in testing but broke in production because the developer had misunderstood a business rule. The existing tool stack made it worse: asset tracking, customs clearance, warehouse management, and last-mile delivery were all handled by separate systems that didn’t communicate.
The Constraints
Non-technical usability was the product. The platform’s commercial premise was that operations managers — not engineers — should be able to build automations. A drag-and-drop interface that still required YAML configuration or API knowledge to actually deploy would not achieve the goal.
Industrial IoT integration. EpicFlow clients run physical equipment — conveyors, forklifts, routing vehicles — that generates operational telemetry. The platform needed to consume that telemetry and act on it, not just connect to software APIs.
Infrastructure neutrality. Some EpicFlow clients operate in environments where data cannot leave the premises. The platform had to support on-premises deployment with the same feature set as the cloud version, without maintaining two separate codebases.
Our Approach
The platform is built around a visual, drag-and-drop workflow designer that exposes complex business process logic as configurable nodes. Connections between nodes define data flow; each node has a business-language interface for configuration, with technical parameters handled by the platform rather than exposed to the user.
Process templates pre-encode common industrial workflows — customs clearance sequences, inventory reconciliation cycles, delivery exception handling — so operations staff start from a working baseline and configure it to their specific rules rather than designing from scratch.
The ML route optimization module integrates with the workflow designer as a callable service. When a delivery workflow triggers a routing decision, the optimizer calculates optimal routes across the fleet considering current traffic, delivery time windows, vehicle load capacity, and driver hours — returning actionable route assignments rather than optimization outputs that require further human interpretation.
IoT integration connects machinery telemetry (vibration, temperature, error codes) to workflow triggers. An anomaly in a conveyor’s vibration pattern triggers a maintenance workflow automatically — creating a work order, notifying the relevant technician, and logging the incident against the equipment record — without human detection or routing.
The platform is containerized via Docker with Kubernetes orchestration, enabling deployment on any cloud provider or on-premises Kubernetes environment from the same build artifact.
The Outcome
- Operations staff now build and deploy automations independently — developer dependency reduced to platform maintenance, not workflow creation
- Fuel and delivery time improved through ML-powered route optimization deployed at scale
- Machinery incidents identified and routed to maintenance automatically through IoT telemetry integration
- Infrastructure flexibility maintained — clients choose cloud or on-premises without feature parity trade-offs
Client feedback: “EpicFlow has transformed how we manage complex logistics operations.” — Emmie Reese, CEO, EpicFlow
Team
Engagement: 9 months, 6 engineers (1 tech lead, 2 backend, 1 frontend, 1 ML, 1 DevOps).
Stack: Angular, RxJS, NgRx, Golang, gRPC, MongoDB, Redis, IBM MQ, Docker, Kubernetes