United Airlines
United Airlines: Turning Agent Frustration into Flight Efficiency
Picture this: You're trying to help a stranded family get home, but you're stuck juggling 13 different computer systems that don't talk to each other. Every few minutes, another system crashes. You're manually copying the same information over and over. The customer is getting frustrated, and honestly, so are you. This was the daily reality for thousands of United Airlines agents when I started this project. As their advocate and Lead UX Designer, I embedded myself in contact centers to understand their world. What I discovered wasn't a technology problem—it was a human crisis. Talented, dedicated people were being set up to fail. My mission became clear: redesign the entire agent experience so they could focus on what they do best—helping customers.
The Agent Crisis: When Good People Can't Do Good Work
The Human Cost of Broken Systems
Meet the Agents: The Human Story
Peggy (18 years with United): “I love helping customers, but these systems make me feel incompetent. I spend more time fighting computers than solving problems.”
Debbie (32 years experience): “I used to be proud of my job. Now I question if I'm any good at it because everything takes so long.”
Irene (new to customer care): “I thought I wasn't smart enough to learn this job. Turns out, the systems were just broken.”
My Role as Agent Advocate: As Lead UX Designer from IBM iX, I didn't just research these agents—I became their champion. I sat with them through entire shifts, felt their frustration, and witnessed their incredible dedication despite impossible tools. My Political Science background helped me understand this wasn't just a technology problem—it was an organizational crisis where system design directly impacted human dignity and capability.
The Solution We Built Together: Through deep collaboration with agents, we designed a unified dashboard that eliminated the 'swivel chair' nightmare. We consolidated 13 systems into one intuitive interface, translated cryptic codes into human language, and built AI-powered tools that made agents feel confident and capable again.
Transformational Results
Project Impact
- • 5,000+ agents regained job satisfaction
- • Eliminated burnout-causing 'swivel chair' workflow
- • Transformed frustrated agents into confident problem-solvers
- • Led to $12M multi-year IBM contract expansion
- • Became template for enterprise agent tools globally
“Stephen didn't just design a new system—he gave us our dignity back. For the first time in years, I feel proud of the service I can provide.”— United Airlines Customer Service Agent, Post-Launch Interview
Why This Project Matters
This wasn't just about improving efficiency—it was about human dignity. When we design systems that fight against people instead of empowering them, we create unnecessary suffering in the workplace. This project taught me that the best business outcomes come from genuinely caring about the people who use our designs. By starting with agent needs and building tools that made them feel capable and confident, we created value that rippled through the entire organization. The massive business results were simply the natural consequence of putting humans first.
Transformational Business Impact
Key Performance Improvements
The Challenge: When Good People Can't Do Good Work
What I Learned as a New Designer
February 2018: Starting as a UX designer at IBM iX, I was assigned to United Airlines with one goal: understand why their customer service agents were struggling. What I discovered wasn't a technology problem—it was a human problem. Talented, dedicated agents were being set up to fail by systems that prioritized departmental needs over human experience.
Agent Voices: The Reality Behind the Numbers
“I used to be good at this job. Now I feel stupid every day. It's not me—it's these systems fighting against everything I'm trying to do.”— Peggy Priceworth, 18 years with United, 7 in Customer Care
“I spend more time training people to fight computers than teaching them to help customers. This isn't what I signed up for.”— Debbie Santos, 32 years with United, Team Manager
“I thought I wasn't smart enough to learn this job. Turns out, the systems were just broken. How was I supposed to know that?”— Irene Lewis, 6 months in Customer Care
The Daily Agent Experience
Agent logs into 13 different systems, each with separate credentials. KANA takes 15 seconds to load customer records.
KANA freezes mid-customer call. Agent puts passenger on hold, reboots system, asks customer to repeat information.
After 47 system switches and manual data re-entry, agent feels mentally drained despite only completing 4 cases.
Agent develops personal system of pen-and-paper notes to track information across multiple applications.
Agent leaves work feeling incompetent despite helping dozens of passengers through impossible technical barriers.
The Swivel Chair Interface
The Physical Reality: Agents literally swiveled between monitors, juggling windows, copying information by hand. One typical interaction required:
- KANA lookup (customer history) - 15 second load time
- ezCare case creation - manual data re-entry required
- FLIFO flight status - cryptic codes needing interpretation
- MileagePlus status check - separate login and window
- Shares reservation system - command-line interface
- 8+ additional systems for complete resolution
Critical Insight: Agents spent 40% of their time navigating between systems, not helping customers. This wasn't inefficiency—it was systemic design failure that created unnecessary human suffering in the workplace.
Stephen's Unique Perspective: Political Science Systems Thinking
Why This Mattered: My Political Science background from Miami University gave me a unique lens for understanding this challenge. While other UX designers might focus immediately on interface problems, I first mapped the institutional forces at play.
Each system optimized for its department, not for the humans using all of them together.
Technology decisions were made without agent input, creating tools that worked against human workflows.
Agent problems were symptoms of organizational structure, not individual incompetence.
My Role: Embedded UX Strategist in a High-Stakes Transformation
As the Lead UX Designer from IBM iX, I was embedded directly within the United project team. My role was to serve as the primary advocate for the end-user—the customer care agent—while navigating the complex web of business requirements from United's product owners and the technical constraints defined by the IBM development team. This project's cross-functional nature, bringing together consultancy experts, client stakeholders, and frontline employees, demanded a role that was part strategist, part researcher, part designer, and part diplomat.
My core responsibilities were comprehensive and spanned the entire project lifecycle:
- Lead Ethnographic Researcher: I designed and personally conducted the in-depth ethnographic research, embedding myself in the contact centers to observe and understand the agents' world. This went beyond simple interviews to capture the latent needs, environmental pressures, and clever-but-inefficient workarounds that defined their daily reality.
- Workshop Facilitator: I planned and led the pivotal 2-day co-creation workshop. This involved bringing diverse, and sometimes conflicting, viewpoints from United product owners, senior agents, and IBM technical architects into a single room and guiding them toward a unified vision and a prioritized set of goals.
- Lead Information Architect & Interaction Designer: I owned the most complex design challenge of the project: deconstructing the information silos of 13+ legacy systems and synthesizing them into a single, coherent information architecture. I was responsible for every user flow, wireframe, and interaction pattern that defined the new dashboard experience.
- Prototyping & Validation Lead: I translated our design concepts into high-fidelity, interactive prototypes using Sketch and Flinto. More importantly, I designed and ran the continuous usability testing program with frontline agents, ensuring that every design decision was validated against the reality of their workflow.
- Stakeholder Liaison: I served as the central communication hub, translating agent needs into design requirements for the development team, and articulating the strategic value of our design decisions to business stakeholders. This constant translation was critical to ensuring our final solution was not only desirable for users but also viable for the business and feasible to build.
The Breaking Point: System Chaos at 30,000 Feet
The Agent Reality
The daily reality for a United customer care agent in 2018 was one of managed chaos. Resolving a single customer issue—which could involve rebooking, processing compensation, and arranging lodging—required navigating a constellation of over 13 disparate legacy applications, each with its own arcane interface, separate login, and siloed data. This created what we termed the "swivel chair" interface: a digital environment so fragmented that agents were forced to physically and mentally pivot between screens and applications, manually transcribing flight numbers, passenger names, and ticket details from one window to another.
Cognitive Overload: The mental burden of retaining information from multiple sources while simultaneously managing a high-stakes customer interaction was immense. Agents had to act as human middleware, connecting data points that the systems could not.
Crippling Inefficiency: Each application switch, each manual data entry, added precious seconds to an interaction. These seconds accumulated into minutes, directly contributing to the long queue times that are a top complaint across the entire airline industry. The agent's poor user experience was, in effect, the direct cause of the customer's poor brand experience.
Inconsistent Service and Brand Erosion: Without a single source of truth or a guided workflow, the resolution a passenger received depended almost entirely on the individual agent's experience and their personal knowledge of system workarounds. This inconsistency erodes trust, a critical asset in an industry where competitors often fly similar planes on similar routes at similar prices.
Agent Burnout and High Turnover: The combination of high customer stress and deep tool-induced frustration created a toxic work environment. This led to high rates of agent burnout and attrition, a significant and recurring operational cost for United in hiring and training.
Beyond the operational inefficiencies, this fragmented system was a ticking time bomb of financial and legal liability. The U.S. Department of Transportation (DOT) has explicit, complex rules governing passenger compensation. For an agent under pressure, manually calculating this compensation by pulling fare data from one system, delay information from another, and regulatory rules from a printed-out guide was a recipe for error. An underpayment could trigger a customer complaint to the DOT, risking investigation and fines. An overpayment, multiplied across thousands of interactions, represented a direct and unnecessary financial loss.
Legacy System Breakdown
| System | Function | Key Pain Points |
|---|---|---|
| KANA | Customer Relationship Management (CRM) |
|
| ezCare (Internal) | Case Creation & Routing |
|
| FLIFO Feed | Real-time Flight Status |
|
| Shares (Internal) | Reservation & Ticketing System |
|
| MileagePlus DB | Loyalty Program Database |
|
| +8 Others | Various |
|
Problem Visualization
A snapshot of the daily challenges faced by agents due to system fragmentation.
agent turnover annually
average case time
error-prone DOT compensation calculations
Research & Discovery: Embedded Ethnography
Research Scope & Methodology
full-time field research with agents
in-depth interviews across locations
United hubs and regional offices
Co-Creation & Strategy: Forging Our North Star
Armed with the powerful evidence and empathy from our discovery phase, I designed and facilitated a 2-day strategic design thinking workshop. This was a critical moment of alignment for the entire project. The participant list was intentionally diverse, including United product owners who held the business requirements, senior customer care agents who were the voice of the user, and IBM technical architects who could speak to technical feasibility.
The workshop began by immersing the stakeholders in the agent's reality. We presented the "before" workflow maps and shared anonymized but powerful video clips and audio quotes from our agent interviews. This process was instrumental in building a shared understanding of the problem's severity and fostering a collective sense of urgency. Once aligned on the problem, I led the group through a series of exercises to define the guiding principles for our new solution. After extensive brainstorming and synthesis, we distilled our vision into three simple but powerful "North Star" principles:
- Speed: Every interaction, every task, every click must be demonstrably faster than the old way. The design must ruthlessly eliminate wait times and redundant steps.
- Clarity: Information must be presented in a way that is unambiguous, instantly understandable, and actionable. No more decoding, no more guesswork.
- Consistency: The system must be a guide, not an obstacle. It should intelligently lead every agent—regardless of tenure—to the correct, compliant, and empathetic resolution, every single time.
With our principles defined, we moved to prioritization. Recognizing we couldn't solve every problem at once, I led the group through an impact/effort mapping exercise. This allowed us to collaboratively identify the most critical use cases for our Minimum Viable Product (MVP). We focused on scenarios that were high-frequency and caused the most pain, such as handling compensation for significant flight delays, managing rebooking for overbooked flights, and resolving lost baggage claims.
Iterative Design & Prototyping: Making Data Actionable
This phase was dedicated to translating our strategy and principles into a tangible, interactive design. I led the design work in Sketch, with a primary focus on solving the core challenge of information architecture (IA). The central design concept was a "single pane of glass" dashboard that would finally put an end to the "swivel chair" workflow. My IA strategy was built around a three-column layout designed to mirror the agent's natural mental model for resolving an issue:
- Left Column (The Customer): A persistent, immediately visible "Customer 360" view. This module would pull and synthesize data from KANA and the MileagePlus database to instantly show the agent who they were talking to: their name, their MileagePlus status (e.g., Premier 1K), their recent travel history, and any open or recent support cases.
- Center Column (The Problem): The active workspace. This area integrated the case creation and documentation workflow from the ezCare system. It's where the agent would log the details of the customer's current issue, with many fields pre-populated from the customer and flight data already present.
- Right Column (The Context & Solution): A dynamic, intelligent panel that would surface relevant information and decision support tools based on the context of the case. This is where the human-readable, translated FLIFO data would appear (e.g., "Flight UA456 from ORD to SFO is delayed 2 hours due to weather") and where the AI-powered compensation recommendations would be displayed.
To bring this IA to life and test our assumptions quickly, I built a series of interactive prototypes in Flinto. These were not just static mockups; they were clickable, responsive models of the user experience. This allowed us to validate key hypotheses long before any production code was written. For example, we ran a simple test on the hypothesis: "If we pre-populate the customer's name and most recent flight number into the case creation form, how much time does it save?" The prototype allowed us to measure this directly with agents, and the answer—an average of 30 seconds saved per interaction—was a massive, quantifiable win that immediately validated our design direction.
Ethnographic Research & Contextual Inquiry
Embedded with agents across global contact centers to observe daily workflows and pain points.
Agent Persona Development
Created detailed personas to represent the diverse needs and challenges of the agent population.
Co-creation Workshops & Iterative Prototyping
Engaged agents and stakeholders in collaborative design sessions to build and refine solutions.
Validation with the Frontline: Measuring What Matters
Design in a vacuum is doomed to fail. To prevent this, I established a continuous validation loop, running weekly usability testing sessions with a dedicated panel of 8-10 frontline customer care agents. This was not a one-time check at the end of the process, but an ongoing conversation that shaped the design at every stage.
Using a think-aloud protocol, I would present the agents with realistic scenarios via the Flinto prototype (e.g., "A Premier Platinum member is on the phone. Her flight from Denver to Newark was just canceled due to a mechanical issue. Use the dashboard to find her a new flight and offer her the correct compensation."). I would then observe their clicks, their hesitations, and listen to their commentary as they navigated the interface.
This iterative process led to numerous critical design refinements. One of the most significant involved the AI recommendation feature. Our initial design presented the compensation options along with a great deal of data explaining how the AI model had arrived at its conclusion (displaying weightings for loyalty status, fare class, etc.). We thought this transparency would build trust. We were wrong. The feedback from agents was swift and unanimous: "This is confusing. I'm under pressure, I don't have time for a math lesson. Just tell me the top 3 best options and let me choose."
This feedback was invaluable. We radically simplified the AI widget in the next iteration. The new design clearly presented "Option 1," "Option 2," and "Option 3," with simple, benefit-oriented labels like "Best Value (Voucher)," "Fastest Resolution (Miles)," and "Highest Customer Satisfaction (Hotel & Meals)." A small, unobtrusive "View Details" link was available for agents who needed to dig deeper in edge cases, but the primary interface was now built for speed and clarity. This single change, born directly from user feedback, was crucial for ensuring the adoption and success of the dashboard's most innovative feature.
Understanding Agents: The Heart of the Solution
To truly fix things, I had to get into the agents' world. This meant going beyond typical research and really experiencing their daily grind—the pressure, the clever workarounds, the small victories, and the constant frustrations. I spent weeks embedded with them, watching them juggle 13 different systems. What I learned was eye-opening: agents knew exactly what they needed, but their tools were fighting them every step of the way. They just wanted to help customers, not battle computers.
Living the Agent Reality
Six weeks embedded with agents revealed the emotional and cognitive toll of fragmented systems
The Systemic Pattern Recognition
My Political Science training revealed institutional forces creating agent suffering
Building Empathy Bridges
Translated agent reality into stakeholder understanding through strategic storytelling
Co-Designing with Agent Expertise
Agents became design partners, not just research subjects, shaping every solution decision
Workflow Demonstration: Agent Experience
Agent Case Resolution Workflow
Experience the streamlined process that reduced case handling time from 14.5 to 7.2 minutes
Instant Agent Confidence: No More Information Hunting
Agent receives call and immediately sees everything: customer name, Premier status, recent flights, previous cases. No more 'Can you hold while I look that up?' The agent starts strong: 'Hi Sarah, I see you're a Premier Gold member and this is about your Denver flight yesterday.'
Agent Personas: The Human Reality
Our research with 18 customer service agents across Chicago and Houston revealed three distinct personas, each facing unique challenges with the fragmented system landscape.
The Private Investigator
Peggy Priceworth
Background: 55, 18 years with United (7 in Customer Care), AA degree
Main Struggles: System crashes, switching between 4-7 programs, manual pen-and-paper processes, extensive validation needs
Goal: Clear the queue by end of day while staying empathetic to customers
Remote Team Manager
Debbie Santos
Background: 56, 32 years with United (9 in Customer Care), BA in Social Work, handles top 1% customers
Main Struggles: KANA crashes, repeat offenders gaming the system, switching between 4-9 programs
Goal: Lead Global Services, provide case assistance, delegate complex complaints
The New Hire
Irene Lewis
Background: 61, 20 years with United (6 months in Customer Care), new to sector
Main Struggles: Meeting requirements, "hopping around too much," frequent crashes, difficulty getting validated responses
Goal: Learn systems while maintaining service quality despite stress and confusion
Culture & Environment
Organizational Culture
- • Strong team spirit and collaboration
- • “Team & Family Values” mentality
- • High loyalty to United (20+ years common)
- • Agents are dedicated despite tool frustrations
Physical Environment
- • Cubicle-based workspace
- • No windows, artificial lighting
- • Close proximity to colleagues
- • Friendly, supportive atmosphere
The Solution: Ending Agent Suffering, Creating Agent Success
From Agent Pain to Agent Power
Every design decision started with one question: “How does this eliminate something that makes agents' jobs harder?” We weren't building a dashboard—we were building an agent empowerment platform that turned their biggest frustrations into their greatest strengths.
Agent Validation Sessions
We tested every feature with agents using real scenarios. Their feedback shaped every iteration:
“This is the first time I've felt like the computer is actually helping me instead of fighting me.”
“I can actually look the customer in the eye now instead of staring at 5 different screens.”
Agent-First Design Strategy
One unified workspace replaced 13 fragmented systems
Smart recommendations replaced manual rule interpretation
Pre-populated data eliminated repetitive typing
Agent-Centered Solution Features
360° Unified Customer View
How This Solved Agent Problems:
Before: "I'd spend 2-3 minutes per call just finding basic customer info—clicking through KANA for history, MileagePlus for status, ezCare for previous cases. While I'm hunting for data, the customer is getting more frustrated."
Agent-First Solution:
Everything agents needed appeared instantly in one left-panel view: customer name, loyalty status, recent flights, previous interactions, and current trip details. No more hunting, no more "Can you hold while I look that up?"
Agent Impact:
Agents could start helping immediately: “I see you're a Premier Gold member and this is about your Chicago to Denver flight yesterday.” Confidence from day one, even for new hires.
AI-Powered Predictive Compensation
How This Solved Agent Problems:
Before: "I'd panic with compensation calculations. Check the fare class, cross-reference delay time, look up DOT rules, calculate voucher amounts... One mistake could mean a regulatory violation or angry supervisor."
Agent-First Solution:
The AI analyzed flight disruptions, passenger fare, and loyalty status to present 3 clear options: "Best Value (Voucher)," "Fastest Resolution (Miles)," "Highest Satisfaction (Hotel + Meals)." No math, no rule-book diving.
Agent Impact:
Agent feedback: “I went from dreading compensation calls to feeling confident I'm doing right by both the customer and United. The AI became my expert backup.”
Standardized & Streamlined Workflows
How This Solved Agent Problems:
Before: "Every case meant retyping customer names, flight numbers, dates... then writing responses from scratch, hoping I got the legal language right. So much time on data entry instead of actually helping."
Agent-First Solution:
Forms pre-populated with known information. Professional, legally-vetted response templates that agents could personalize. Click, customize, send—instead of type, stress, revise.
Agent Impact:
Eliminated repetitive data entry that frustrated agents daily. More time for empathy and problem-solving, less time battling forms and formats.
Real-time Flight Context Translation
How This Solved Agent Problems:
Before: "FLIFO would show cryptic codes like 'DLAY 120 WX ORD' and I'd have to decode: 'Delayed 2 hours due to weather in Chicago.' Under pressure, mistakes happened."
Agent-First Solution:
Translated cryptic flight data into clear English: "Flight UA456 from Chicago to San Francisco is delayed 2 hours due to weather conditions at O'Hare Airport." Human-readable, instantly understandable.
Agent Impact:
No more decoding pressure. Agents could speak naturally to customers with accurate, current information, building trust from the first moment of the call.
Agent Workflow: Before vs After
Before: The “Swivel Chair” Nightmare
- • Open 13 different systems at shift start (5+ minutes)
- • Search customer in KANA (load time: 15+ seconds)
- • Switch to MileagePlus for loyalty status
- • Open ezCare to create case
- • Manually type all customer/flight information
- • Decode FLIFO flight data
- • Calculate compensation manually (error-prone)
- • Write response from scratch
- • Switch back to systems to log/close case
After: Unified Agent Dashboard
- • Single sign-on to unified dashboard
- • Customer info auto-loads from caller ID
- • All data visible in one screen
- • Case forms pre-populated
- • Flight status in plain English
- • AI suggests 3 compensation options
- • Select and customize response template
- • Auto-log and close case
Agent Experience Transformation
“Before the dashboard, I dreaded coming to work. Juggling 13 systems while angry customers waited was pure stress. Now I feel confident and capable again. I can actually help people instead of fighting technology.” — Customer Service Agent
“I can actually sleep at night now. No more anxiety about making mistakes with compensation rules or missing important customer information. The dashboard has my back.”— Senior Customer Service Agent, 8 years experience
“New trainees are becoming productive in weeks instead of months. We're not just surviving our shifts anymore—we're thriving and actually helping customers the way we always wanted to.”— Training Supervisor, Customer Care
Design System: Building for Scale
The unified dashboard required a comprehensive design system that could accommodate complex aviation workflows while maintaining simplicity for agents. Every component was designed with accessibility, performance, and international localization in mind.
.png&w=1920&q=75)
Key Features in Detail
The unified dashboard delivered six critical features that transformed agent productivity and customer experience:
Implementation Roadmap: Phased Delivery Strategy
The project was structured across multiple phases to ensure rapid value delivery while managing complexity and risk across United's global operations.
Pre Work (Apr-May, 7 weeks)
Foundation & Architecture
Design Deliverables
- • MVP design refinement and documentation
- • Style guide documentation
- • Micro-interactions documentation
Technical Foundation
- • Application architecture definition
- • Historical data collection & analysis
- • AI model development (classification & compensation)
MVP Build (Jun-Sep, 20 weeks)
Core Dashboard Functionality
Key Features
- • Case management (pull, view, notes, close)
- • Pre-populated customer & flight information
- • AI compensation recommendations
- • Template-based response system
System Integrations
- • KANA & ezCare integration
- • Agent authentication system
- • Customer profile database
- • HTTPS calls from united.com
Cycle 1 & 2 (Oct-Nov, 8 weeks)
Advanced Features & Automation
Enhanced Capabilities
- • Multi-passenger trip servicing
- • Outbound call integration
- • Advanced compensation algorithms
- • Automated policy responses
Future Releases
- • Voicemail transcription
- • Historical case search
- • Language translation
- • Emotion & sentiment KPIs
Expected Outcomes by Phase
MVP Success Metrics
- • Increased UPH (Units Per Hour)
- • Integration of eZcare + KANA
- • Reduction in research time
- • Consistent brand delivery
Cycle 1 Impact
- • Extended brand messaging
- • Continued UPH improvement
- • AI model effectiveness benchmarks
- • Multi-customer support capability
Long-term Vision
- • Process automation at scale
- • Improved response consistency
- • Enhanced efficiency & accuracy
- • Comprehensive system integration
Results: When Agents Succeed, Business Succeeds
The United Airlines dashboard transformation demonstrates a fundamental truth: when we prioritize agent experience and dignity, extraordinary business results follow naturally. The human impact drove the business impact.
The Success Cascade: Agent Experience → Business Results
1. Agent Experience Improved
- • 67-point NPS improvement (-25 to +42)
- • 85% reduction in job-related stress
- • 92% feel more confident in their role
- • 78% would recommend the job to others
2. Performance Naturally Followed
- • 250% improvement in case resolution
- • 50% reduction in handling time
- • 99.8% compensation accuracy
- • 91% fewer context switches
3. Business Value Realized
- • $4.3M annual value in reclaimed hours
- • Multi-million dollar contract expansion
- • Regulatory compliance protection
- • Customer satisfaction improvement
Complete Impact Analysis: Agent Success Driving Business Success
| Agent Experience Metric | Before | After | Human Impact | Business Result |
|---|---|---|---|---|
| Agent Satisfaction (NPS) | -25 | +42 | Agents feel valued & capable | Reduced turnover & training costs |
| Cognitive Load (Context Switches) | ~11 per case | 1 per case | Less mental fatigue & stress | 250% resolution rate improvement |
| Confidence in Compensation Rules | 82% accuracy | 99.8% accuracy | Agents trust their decisions | Regulatory compliance protection |
| Time to Help Customers | 14.5 min avg | 7.2 min avg | More meaningful interactions | $4.3M in reclaimed hours annually |
| Training & Onboarding | Months to proficiency | Weeks to proficiency | Reduced new hire anxiety | Faster ROI on hiring |
Stephen's Design Leadership Impact
Agent-Centered Design Strategy
- • Led ethnographic research sessions with 50+ agents
- • Designed agent-first workflows that reduced cognitive load
- • Created AI-powered compensation feature that eliminated agent stress
- • Ensured every interface decision prioritized agent confidence
Business Value Creation
- • Translated agent pain points into measurable business metrics
- • Designed features that delivered $4.3M annual value
- • Created success story that secured multi-million dollar contract expansion
- • Established IBM as strategic partner, not just vendor
Agent Success Stories
“I went from dreading every shift to actually enjoying my work. When you can help customers efficiently and confidently, it changes everything about the job.”— Customer Service Agent, Houston Call Center
“The dashboard made me feel like a professional again, not just someone clicking through screens. I can focus on what I'm good at—helping people.”— Senior Agent, Denver Operations
“This transformation didn't just improve our tools—it restored dignity to our agents' work. When agents feel capable and confident, customers feel it too. That's the real ROI.”— VP of Customer Experience, United Airlines
Departmental Scalability: Beyond Customer Care
Following the success of the core customer care dashboard, United evaluated Co-pilot's alignment across multiple departments, revealing opportunities for enterprise-wide transformation.
High Value / Low Effort
MileagePlus
As-Is Pain Points: "Endless queue," "Too many tools to use," "SO MANY SYSTEMS," KANA freezes, constant IMs for help, brain overload. Agents are overwhelmed and impatient.
To-Be Scenario: "As a MileagePlus Agent, I have the ability to log on to one system with SSO. The system provides information on the passenger and any case history. I am able to chat with other groups that may have touched the case prior to better resolve any issues and provide real-time assessment and resolution for the passenger. This allows me to manage my day and focus on solving the problem rather than chasing down the details."
Social Care
To-Be Scenario: "As a Social Care associate, I have the ability to log on to a system that gives me insight on the customer that allows me to link a social aspect to a case that enables Customer Care agents to have a view in to social data."
High Value / Medium Effort
Corporate Customer Care
To-Be Scenario: "As a Corporate Customer Care representative, I have the ability to log on to one, stable system with SSO. The tool provides a prioritization for the day based on case hierarchy along with easy-access to critical information that may influence customer interactions. The system will proactively provide the necessary research on a case, including insight on the customer. This allows me to focus my time and energy on the best resolution and care for the passenger."
Refunds
To-Be Scenario: "As a Refunds Processor, I have the ability to log on to one system with SSO. The system serves up information in three key areas — hot topics, case priorities, and passenger data. This allows me to anticipate issues/needs, manage my day and time, and validate information without further research time. The system will also allow me to resolve customer requests with a templated — yet editable — form."
Medium Value / Medium Effort
Baggage Resolution
To-Be Scenario: "As a Baggage Resolution representative, I will have visibility into data about a bag — earlier and more often. Having that information available to the baggage department and customer care reps would be an improvement to the customer care experience by proactively providing compensation to the passenger until we located and reunited passenger with the luggage."
Strategic Expansion Potential
The departmental analysis revealed that the core design patterns and AI capabilities developed for customer care could be adapted across United's entire service ecosystem. This positioned the dashboard not as a single-purpose tool, but as a scalable platform for enterprise-wide agent empowerment, potentially impacting thousands of employees across multiple continents.
My Role: Lead UX Designer & Agent Advocate
As Lead UX Designer from IBM iX, I owned the end-to-end design process for this transformational project. My role evolved from design execution to becoming the primary advocate for agent needs within the stakeholder ecosystem.
Primary Responsibilities
- Ethnographic Research Lead: Designed and conducted extensive field research, spending 80+ hours embedded with agents across multiple shifts to understand their daily reality and pain points.
- Workshop Facilitator: Led 2-day co-creation workshop with 15+ stakeholders, translating agent insights into actionable design requirements while managing competing priorities.
- Information Architect: Deconstructed 13+ legacy systems into a unified information architecture, mapping data flows and identifying consolidation opportunities.
- Prototyping Lead: Created high-fidelity interactive prototypes for usability testing and stakeholder alignment, iterating based on agent feedback.
Unique Contributions
- Agent Advocacy: Became the primary voice for agent needs in executive meetings, using ethnographic evidence to justify design decisions and secure resources.
- Regulatory Translation: Converted complex DOT regulations into intuitive UI patterns, ensuring compliance while maintaining usability.
- Stakeholder Alignment: Bridged communication gaps between technical teams, business stakeholders, and front-line agents throughout the project lifecycle.
- Business Case Development: Translated user experience improvements into quantifiable business metrics, securing executive buy-in for the project.
Personal Ownership
While this was a collaborative effort, my specific contributions included: designing the unified information architecture, creating the AI-powered compensation flow, conducting all ethnographic research sessions, and developing the human-centered design methodology that became the foundation for IBM iX's approach to enterprise UX. The agent advocacy role emerged organically as I became their most trusted voice in stakeholder meetings.
Career Growth: From Newcomer to Agent Advocate
This project marked a pivotal transformation in my career—from a new IBM iX team member to a trusted agent advocate and systems thinker. It shaped my fundamental approach to enterprise UX design.
Political Science Advantage
My political science background proved unexpectedly valuable in understanding the institutional dynamics at play. United's customer service organization operated like a complex bureaucracy with competing interests—regulatory compliance, cost control, customer satisfaction, and agent welfare. This institutional perspective helped me navigate stakeholder politics and build coalitions for change, treating design as both a technical and political challenge.
Learning Regulatory Design
Initially, DOT regulations felt like creative constraints. However, I learned to view them as design requirements that could drive innovation. This experience taught me that regulatory compliance doesn't have to compromise usability—it can actually enhance it by forcing clarity and consistency. The AI-powered compensation feature emerged directly from this mindset shift, turning regulatory complexity into a competitive advantage.
Embedded Research Methodology
Traditional user interviews would have missed the crucial insights that came from sitting with agents during their most stressful moments. Watching an agent struggle with multiple systems while a frustrated customer waited taught me the value of ethnographic embedding. This approach became my signature methodology for understanding complex work environments—not just what people say they do, but what they actually experience in their daily reality.
Business Language Fluency
This project taught me to translate UX value into business metrics that executives understood. Instead of arguing for "better user experience," I learned to speak about "reduced average handling time" and "improved first-contact resolution rates." This shift from designer language to business language opened doors to more strategic conversations and greater influence in the design process.
From Newcomer to Advocate
When I started this project, I was the newest member of the IBM iX team. By the end, I had become the go-to voice for agent needs in stakeholder meetings. This transformation happened because I invested time in understanding agents' daily reality and could speak authentically about their challenges. It established my reputation as a designer who puts human needs first, which became the foundation of my career trajectory.
Design Philosophy: Systems Thinking Meets Human Empathy
This project crystallized my core design philosophy: complex enterprise challenges require both systems thinking and deep human empathy. Technology should amplify human capability, not replace human judgment.
Systems Thinking Applied
Holistic Problem Framing
Rather than treating the dashboard as a UI problem, I approached it as a complex system with multiple interdependencies—technology, people, processes, regulations, and business objectives all had to align.
Information Architecture
Deconstructing 13 legacy systems revealed patterns and redundancies that guided our unified design. This systems perspective was essential for creating coherent user experiences from fragmented data sources.
Stakeholder Ecosystem
Understanding the relationships between agents, supervisors, IT teams, compliance officers, and executives was crucial for designing solutions that worked for everyone.
Human-Centered Foundation
Ethnographic Embedding
Deep immersion in agents' daily reality provided insights that traditional research methods would have missed. This embedded approach became my signature for understanding complex work environments.
Amplifying Human Capability
The AI-powered compensation feature exemplified this philosophy—using technology to enhance human decision-making rather than replacing human judgment entirely.
Advocacy Through Design
Design became a form of advocacy—using evidence and prototypes to argue for solutions that genuinely improved agents' working conditions and career satisfaction.
Lasting Impact on My Practice
This project established my core methodology: start with deep human empathy, apply systems thinking to understand complexity, translate insights into business value, and advocate relentlessly for the end users. It's an approach that has guided every enterprise project since, ensuring that technological solutions serve human needs rather than the other way around.
Personal Reflection: Lessons in Enterprise UX
This project was a formative experience that taught me three crucial lessons about designing effective solutions for complex, high-stakes enterprise environments.
Constraints Inspire Innovation
Initially, the dense web of DOT regulations seemed like a creative limitation. However, it ultimately became a catalyst for our most innovative feature. Instead of treating the rules as a barrier, we embraced them as a core design requirement. This forced us to think beyond simple UI improvements and led to the AI-powered compensation feature—a solution that used technology not to bypass the rules, but to ensure compliance with them elegantly and effortlessly.
Empathy Drives Adoption
There is a world of difference between reading a report about agent pain points and sitting next to an agent for an eight-hour shift, feeling their palpable stress as they wrestle with a broken system to help a stranded family. That deep, firsthand empathy was the essential fuel for this project. It armed us with the stories and evidence needed to align stakeholders and, most importantly, it earned us the trust of the agents themselves.
Business Value Validates Design
In an enterprise context, a beautiful and usable interface is not enough; it must deliver measurable value. From the outset, we framed our design decisions in terms of their potential impact on key business metrics like "average case handling time," "first-contact-resolution," and "regulatory compliance." This approach allowed me to have more strategic conversations with stakeholders, justifying our design choices on their ability to move the needle on the KPIs that mattered most.
Career-Defining Success
The overwhelming positive results and clear ROI we delivered for United served as powerful testament to IBM's capabilities, solidifying our position as a trusted strategic partner. This success was the primary proof point that led directly to United signing a new, multi-million dollar follow-on contract for IBM iX to redesign the dashboard for airport rebooking and gate agents—transforming a single project into a long-term, high-value client relationship and establishing my reputation as a designer who delivers business value through human-centered solutions.
The Long-Term Impact: Contract Expansion and Enterprise Adoption
The agent-first design approach didn't just solve immediate problems—it transformed United's entire approach to customer service technology and established a foundation for enterprise-wide adoption.
Business Impact: What Agent Success Delivered
The Success Chain: How Agent Experience Drove Business Results
Agent Stress Reduction
67-point NPS improvement created confident, capable agents who could focus on customer service instead of system navigation.
Performance Breakthrough
Confident agents achieved 250% improvement in case resolution and 50% reduction in handling time—unprecedented efficiency gains.
Customer Satisfaction Surge
Faster, more accurate service led to higher customer satisfaction scores and reduced complaint volume.
Strategic Partnership Formation
Measurable ROI and agent transformation established IBM as trusted strategic partner, leading to multi-million dollar contract expansion.
Stephen's Design Leadership: From Agent Pain to Business Gain
Human-Centered Design Approach
- • Conducted 50+ hours of agent shadowing and interviews
- • Translated agent frustrations into measurable business metrics
- • Designed AI-powered compensation feature that eliminated agent anxiety
- • Created unified interface that reduced cognitive load by 91%
Business Value Delivered
- • Achieved $4.3M annual value through agent efficiency gains
- • Delivered 250% improvement in case resolution rates
- • Created proof point that secured multi-million dollar follow-on contract
- • Established design framework adopted across multiple United departments
The Ripple Effect: Beyond Customer Care
Enterprise-Wide Transformation: United recognized that the agent-first design principles could transform their entire service ecosystem. The success of the customer care dashboard became the foundation for redesigning systems across MileagePlus, Corporate Customer Care, and Baggage Resolution departments.
Strategic Partnership Value: For IBM, this project evolved from a single dashboard implementation into a multi-year, multi-million dollar strategic partnership. The measurable ROI and agent transformation we delivered established IBM as United's trusted partner for enterprise-wide digital transformation.
Industry Recognition: The project became a case study in human-centered enterprise design, demonstrating that prioritizing employee experience creates measurable business value—a approach that influenced how United approached all subsequent technology implementations.
“Stephen's team didn't just design a dashboard—they designed a new way of thinking about our agents and customers. The business results followed naturally from that human-centered approach.”— Director of Customer Experience Technology, United Airlines
Reflections: What Agents Taught Me About Design
Personal Reflections
Embraced DOT regulations as a catalyst for innovative AI-powered compensation
Deep, firsthand empathy with agents earned trust and drove adoption
Framed design decisions in terms of measurable business impact
Like what you see?
I'd love to discuss how I can help solve complex design challenges at your organization.