Background

Client: 

Carnegie Technologies, Inc. Internal Product

Request: 

Carnegie Technologies had already built a partial platform—a web application that presented and parsed data captured by LoRa‑based sensors for a range of industrial and commercial use cases. I joined to review and improve usability and information architecture, as well as design additional features required to meet the established business requirements.

Challenges: 

• Two years of work had already gone into the project, yet there were fundamental problems from the information architecture to the business strategy.

• Usability and hierarchy issues plagued the existing design, making it hard to maintain consistency across the UI and user flows.

• The business was caught in a sunk‑cost fallacy—continuing to patch a fundamentally flawed design instead of applying what we’d learned to build a new foundation.

• New requirements kept being added to the prioritization queue before the original MVP could be delivered.

Discovery

As a new employee, I got up to speed quickly on the product’s intent while still learning its current state. Beyond grasping the product itself, I had to understand how Carnegie’s Agile process worked. This multi‑pronged discovery required me to run several parallel exploration efforts.

UX audit

It didn’t take long to realize that the Longview application was plagued by numerous fundamental architectural and usability problems—issues that stemmed from inadequate design processes and vetting.

My first task was to conduct a UX audit, identifying both short‑term fixes and longer‑term strategic changes needed to improve overall functionality.

Short term suggestions from the audit included:

• Adopt a more consistent color palette across the application

• Remove “functional controls” from the navigation menu to avoid buttons serving dual purposes

• Standardize button states globally (normal, hover, active, disabled)

• Enhance accessibility by using icons and shapes in addition to color cues

• Redesign map item visuals to reduce visual noise and prevent overlapping markers.

Long term suggestions from the audit included:

• Investigate mobile usage patterns and trim functionality accordingly

• Re‑organize the information hierarchy, nesting related features and sections more logically

• Revise the navigation structure so users can intuitively

• locate their current “location” in the app

• Strengthen alert handling and management for clearer user communication

• Evaluate building a new version from scratch to resolve core information architecture and structural problems.

Getting up to speed

Stepping into a project that had already been underway for several years presented many challenges. Working with the team, we gradually introduced better design and development processes and identified the critical areas that required further refinement.

Because the original information architecture approach was fundamentally flawed, I proposed refactoring large portions of the interface. However, the business was hesitant to start from scratch, so I focused on bridging those gaps while establishing a workshop process for delivering newly requested functionality.

Design

Because the project was already underway and the business wanted to focus on building out new features, features were prioritized and tackled one at a time. This essentially resulted in several mini-projects focused on each new feature.

Designing new features

Over the next year I collaborated with business owners and product managers to define each new set of requirements—often through alignment workshops—before producing wireframes and designs for review and implementation. Some of these efforts are outlined and described below.

Location history interface

To give users the ability to review historical sensor data—both in a table and on a map—I designed several interfaces. The history can be accessed from the Details pages (tabular view or a historical track) and also directly from the Map view.

Functional requirements:

• Specify a time range for the history

• View the historical track on the map

• Download reports

• Display the timestamp, reading location, and coordinates.

Assets creation interface

Users needed the ability to create “assets” that could have sensors attached. Until the interface and workflow were complete, the backend team had to create assets manually. I focused on ordering the steps so users avoided invalid choices while still allowing optional fields to be skipped.

Workflow:

• Select an asset category

• Enter required information

• Choose sensors to assign

• Edit selected sensors (if needed)

• Receive a completion toast and option to create another asset.

After delivering the initially requested features, I convinced the business stakeholders to let us begin designing the next platform version with a more structured process. Seizing that opportunity, I applied my Human‑Centered Design training to foster a truly collaborative approach.

Human-Centered Design workshops

I designed and facilitated a series of workshops with cross‑functional team members to capture desired outcomes for the new version. The items were grouped into Affinity Clusters for categorization, after which the team collaborated on an Importance/Difficulty Matrix exercise. This shared environment helped everyone grasp the scope that remained and fostered interdisciplinary insight into how tasks were prioritized—and why.

Increasing agility

I also took the initiative to resolve workflow problems with the team. Recognizing that a hybrid agile approach was unsuited for a project of this scale, I began guiding the group toward a more agile process. To support this transition, I designed a set of magnetic whiteboard “components” that could be instantly dropped onto sketches. A companion component library enabled the development team to move directly from those quick wireframes to functional prototypes—facilitating rapid iteration and continuous improvement.

Building with components

Collaborating with the dev team, we successfully converted several legacy elements into reusable components—identified during the whiteboard wireframing sessions. New interface designs were rapidly assembled from these components, first as prototypes and then refined through UX and QA testing to confirm that functionality, usability, error validation, and warning feedback met our standards.

The interfaces shown below result from the sketches above built with pre-made developer components.

Results

Although I departed Carnegie before the Longview application reached completion, substantial progress had been made on implementing the required functionality. Along the way, I helped institute new processes that replaced exhaustive documentation with stronger collaboration, visibility, accountability, and ownership across a previously siloed team. Additionally, I identified several overlooked areas outside the application itself—such as online hardware/software sales, customer support cycles, and end‑of‑life planning—that were critical to the overall customer journey.

My Role: UX Product Design Lead

Design Team:

• Thomas Brady (UX Team Lead)

• Mick Santostefano (Prototype Design/Coding)

• Nathan Dominguez (Design/Wireframe Production Support)

Client: Carnegie Technologies, Inc. Internal Product

Responsibilities:

• User Experience Strategy

• Creative Direction

• Process Design

• Requirements Documentation

• UI Design

Postscript: A prototype

Early in the project, I directed the creation of a quick interactive prototype to steer the organization toward a new design paradigm—using the existing assumptions and requirements as a starting point while explicitly flagging that those premises should be reexamined. The visual language was refreshed for modernity and legibility, and we proposed a substantial overhaul of the information architecture. Although this approach was built on uncertain assumptions, it aimed to deliver a more marketable and usable product.

The leadership team appreciated the cleaner design but remained hesitant to abandon work on the original version. Consequently, the prototype direction was shelved.