Gregory Pearson . Design

44.95°N, 93.09°W

(UTC)

Gregory Pearson . Design

44.95°N, 93.09°W

(UTC)

AI Empowered Front Line

CATEGORY

Case Studies

Client

Exelon Corporation

DATES

2025—Current

Exelon is the nation’s largest energy provider, serving more than 10 million customers across several operating companies. The scale is not noted as a design accolade, but as context for the weight their internal tools carry. This project involves building new AI powered systems to support their front line. These are the customer service representatives who work under pressure and rarely with pause. The tools were designed to reduce noise, create calm, meet the pace of live calls, and give them space to do their best work.

Opportunity ⇢

Build systems that reduce noise, create calm, and accelerate work, giving time, clarity, and confidence back to the front line teams who keep the customer experience moving.

That direction came into focus after engaging with leadership and a wide range of representatives across the organization. The needs were consistent and clear: less mental load, fewer obstacles, and tools that support flow instead of interrupting it.

This effort is exceeding enterprise goals for efficiency, accuracy, and consistency, in both how the tools were built and how they perform in use. More importantly, they have earned genuine trust. A parallel goal was to model how adoption can grow when tools are positioned as partners, accelerators that augment the unique talents of the people who use them.

Context

On the front line, dedicated teams of customer service representatives handle fast moving, high pressure work surrounded by layers of systems, screens, and shifting information. Each call can involve multiple tools, dozens of data points, and a level of precision that leaves no room for error.

And time? Well, time is scarce: on average, there are a whole four seconds between calls. In that kind of environment, focus is everything. These tools had to respect that reality, managing complexity in the background so representatives could stay clear, confident, and present.

Role

My focus has been defining how AI should function in this environment. I lead product design across three connected tools: Knowledge Base, Email Assist, and Call Assist. Together they now form a unified AI Platform.

I have also been given the opportunity to help shape Exelon’s broader approach to AI. The discipline behind this work, the way we surfaced real needs, and the patterns that emerged are now influencing how other teams decide where and how AI should be applied.

Interaction DETAILS THAT signal presence

FEEDBACK sessions shape how each tool evolveS

Framing

At its core, the design philosophy was about restraint. These tools were built to evolve with the people and systems they support. Every decision traced back to a single question: will this help our users stay focused?

That led to interfaces that speak softly, interactions that happen without fanfare, and patterns that stay consistent across every surface. The goal was not to replace effort with automation, but to amplify it.


Strategy

The broader strategy was to bring AI to the front line in a way that felt personal, practical, and measurable. That meant designing experiences that fit existing workflows instead of redefining them and structuring feedback systems that revealed how real users interacted with each feature before scaling. These tools became a learning system as much as a product, designed to adapt and improve with each interaction.

Under the surface, each product is powered by retrieval augmented generation architecture, combining vector search, prompt orchestration, and secure model routing. This allows the system to surface reliable, context aware responses within the strict compliance and security requirements of a regulated enterprise.

Execution ⇢

Knowledge Base became the foundation, the trusted source for fast, reliable information. Email Assist extended that foundation to written communication, providing context aware drafting support for outbound responses. Call Assist carried it forward into live conversations, generating real time transcripts and clear summaries once the call ended.

Each product carried the same DNA: simple interfaces, minimal intrusion, and clear, reliable output. Every interaction was tested with real users, refined through observation, and aligned with engineering to match the pace and complexity of the work.

The design system expanded to include new AI specific components, motion behaviors, and documentation patterns, each built from the same foundational atoms to maintain cohesion across the enterprise.

Built for flow.

Tools that anticipate needs and stay out of the way.

retrievING the right data in real time

Defining A shared "AI PLATFORM"

Expanding the enterprise design system

Impact

This quiet but powerful suite of tools has formed a new backbone for Exelon’s customer care. They reduce friction, increase accuracy, and create space for focus where time is scarce. Each tool listens first, acts second, and evolves responsibly as part of a unified system that scales with the organization.

This effort is exceeding strategic objectives and helping advance Exelon’s long term vision for operational efficiency and responsible AI adoption.

A parallel effort in this work has been shaping Exelon’s broader AI strategy. The process used here has become a model for how teams diagnose the true need, ground decisions in evidence, and identify where AI can make a meaningful difference. That pattern is now guiding other AI efforts across the organization.


Reflection

The value of this project came from listening closely, building with evidence, and shaping tools that respect the rhythm of their work.

These tools resonate because they empower.

They amplify the unique skill sets of the front line teams, and they stay quiet in the moments that matter.

MORE Case Studies

Loading

Modernizing Internal Systems

2024—2025

Loading

Tracker for Securian

Loading

ClearRate for Securian

Loading

Modernizing Internal Systems

2024—2025

Loading

Tracker for Securian

Loading

Modernizing Internal Systems

2024—2025

Loading

Tracker for Securian