Work
Logistics / Operational Intelligence

Operational Intelligence: From Data Entry to Decision Support

How we re-architected a legacy logistics system to support complex decision-making and reduce cognitive load.

The Problem (The Legacy Friction)

The legacy system was a "High-Density Utility" prioritizing raw data entry speed over human learnability. It required dispatchers to manually translate real-world logistics into rigid database rows.

Recall over Recognition

The interface relied entirely on memorized shortcuts. With 25% of documentation dedicated to "F-Keys" (F2 to Submit, F5 to Rate), the system prioritized active recall over visual recognition, creating a steep "memorization cliff" for new hires.

The "Logic Knot" of Routing

The system treated routes as flat, linear arrays. To manage a commingled route, dispatchers had to manually shuffle stops using "Move Left" and "Move Right" buttons. This manual array manipulation was the root cause of sequencing errors.

Hidden Constraints

Critical logic was hidden behind "mystery meat" navigation. Vital constraints like Hazmat requirements were indicated only by subtle blue outlines on buttons, forcing users to constantly scan for "invisible" signals.

The Solution (The Architect’s Fix)

We didn't just "skin" the legacy Java application; we re-architected the mental model of logistics dispatch.

Solving the Commingled Route Paradox

The legacy system forced dispatchers to manage complex, multi-customer routes as flat, linear lists using manual "Move Left/Right" buttons. We introduced a Hierarchical Logic Visualization. The new UI groups orders visually by Route → Customer → Stop, allowing dispatchers to drag-and-drop complex itineraries. This offloaded the "mental mapping" from the human brain to the interface.

Reactive Logic vs. Static Input

We moved from a "Recall-based" system (memorizing F-keys like F5 to rate) to a Reactive State Model. Pricing, route time estimation, and error validation now calculate instantly as data is entered. We replaced hidden constraints (subtle blue outlines) with Contextual Guards—high-visibility alerts that block invalid states before they are committed.

The "Glass Box" Transition

We migrated the rigid "Single Screen Density" layout into a Progressive Disclosure workflow. Information is now contextual—users see "Hazmat Details" only when relevant, reducing visual noise by ~40% while maintaining high data density for power users.

The Outcome (The Evidence)

The redesign transformed the Order Entry system from a passive database terminal into an active strategic tool.

Reduced Cognitive Load

By visualizing the route hierarchy, we eliminated the mental math required to understand commingled routes, significantly reducing dispatch errors on multi-stop orders.

Accelerated Onboarding

We shifted from "Memorization" (F-keys) to "Recognition" (UI Controls). New CSRs and dispatchers achieved proficiency in days rather than weeks, as the system now guides them through the "Happy Path."

Operational Velocity

The shift to a web-based, reactive architecture allowed for real-time updates without page refreshes, increasing the orders-processed-per-hour metric for high-volume centers.

Let’s Talk About Your Work

If you’re facing similar challenges—or want to understand how this approach could apply to your product—we’re happy to talk.

Start a conversation