Redesigning AT&T's Growth Engine
Role: Lead Product Designer, owning the process end-to-end across Product, Marketing, Engineering, and Design Systems, hands-on with the team at every stage, partnering with 45+ stakeholders across 8 product scrum teams.
Research & Discovery: 2.9M Visitors Mapped
Working cross-functionally with Research, Product, and Engineering, we mapped the funnel behind that drop-off: 2.9M visitors entered, most never finished. I led the design sprint in a cross-collaborative effort, synthesizing personas alongside the research team. Five personas ensured the redesign worked beyond the average user, refined through 5 testing rounds and 400+ surveys.
Design System
I helped scale and drive adoption of a shared design system across the upgrade flow, building new components that became reusable across the platform.
Core system elements: grid, color, typography, and components.
Testing Assumptions: Balancing User and Business Needs
Business needed upsell and trade-in revenue. Customers needed speed and clarity, two goals that don't naturally align. Rather than treating that as a fixed trade-off, I used survey data to make the case for where they could coexist: research showed upselling mid-flow drove drop-off, so I moved revenue goals to after checkout instead of competing with the core task. Data, not opinion, aligned both sides.
Feasibility shaped the design too. As an example, real-time price updates across 8 steps required close collaboration with engineering, I drove that collaboration to keep performance and accuracy in sync.
Optimizing Conversion: 23 Steps to 8
An editable, single-canvas approach reduced the rigid 23-step sequence to 8 steps, cutting cognitive load. It also let users configure, edit their choices, and pick up across devices without losing progress.
The flow starts personalized, by name and past purchases, guided by a stepper that shows exactly where the user is. Live pricing updates at every choice, ending in a consolidated recap at checkout, an end-to-end experience across configuration, trade-in, and payment.
Accessible by Design
Partnering with the accessibility team, we improved color contrast, larger touch targets, and plain language, so the redesign worked for every customer, not just the most tech-comfortable ones.
Doubled Conversion in 3 Months
Speed to market: rolled out in three months (previously was18 months), this drove a 230% spike in online upgrades, setting a new company record.
This approach became AT&T's model for checkout flows shop-wide. Post-launch, we tracked conversion and drop-off weekly to confirm the gains held, not just at rollout, giving the team a fast signal if any step regressed.
What I'd Do Differently
The redesign personalized content, like name and past purchases, but every user still moved through the same steps. Today, I'd have the flow notice what users want from the time they pause and when they backtrack, then adjust which steps and offers they see: a premium path for a confident user, financing surfaced when someone hesitates. I'd also add an AI assistant option: a user types "best camera under $30/mo" and the flow configures around that, showing the phone with the strongest camera in that price range instead of a generic list.
I'd resolve credit eligibility too, surfaced at step one instead of after configuration. Tradeoff: harder to test across segments, so I'd pair it with clear guardrails on what can vary.
