Millions of people share a wireless plan. Almost nobody shares the bill gracefully — it's screenshots, Venmo requests, and awkward group texts. I led UX for the industry's first carrier-native bill-splitting experience, turning a monthly chore into two taps and a text message.
What I ownedI led two designers. I owned the design direction and mapped every flow and edge case, took the work to Business, Product and senior leadership for approval, and made the design go/no-go call. The designers built out the experience from that frame.
Money collected through SplitPay that previously moved outside AT&T — through cash, third-party apps, and group texts.
Doubled against the launch baseline. Twice as many assigned payers finished paying their share after the flow was simplified.
Fewer calls to care each year, because account holders stopped phoning in to sort out who owed what.
Net new adds attributed to SplitPay — people added lines once sharing the bill finally worked.
Families and friends pool onto one account to get multi-line discounts. Then, every single month, one person does the math, chases four people, fronts the money, and hopes nobody's line gets suspended.
Before SplitPay, that job lived entirely outside AT&T — in spreadsheets, third-party payment apps, and group chats. Billing was already one of the largest friction points in the customer journey, and this was friction we were quietly outsourcing to our customers.
The design bet: if AT&T could split the bill at the source and let each person pay their own share directly, we'd remove the awkwardness, protect the account, and give people a reason to add lines instead of dropping them.
Account holders hand-calculated line-level charges, taxes, and installment plans to figure out who owed what.
Late fees and suspension hit the whole account — so the owner absorbed the risk for everyone else.
19% of eligible customers offered SplitPay in CX Labs expressed interest; 32% of those who requested more info went on to enroll.
SplitPay had no precedent to copy — no competitor had shipped it. That meant the design work wasn't just screens; it was defining the mental model, the vocabulary, and the rules of a product that didn't exist yet, then keeping seven teams aligned on it across two years.
Owned the end-to-end UX vision across enrollment, payer assignment, payment status, settings, and cancellation — including the prototype that got the concept funded.
Turned brutal billing-system edge cases — mid-cycle changes, cancelled lines, wearables that can't pay for themselves — into states a normal human can understand.
Ran design reviews spanning Product, Business, Engineering, Content, CX, Experimentation, Support and the App team — turning conflicting asks into one shippable direction.
Led two designers through day-to-day execution — daily working sessions, critiques, and Design Standards reviews that kept quality high without gatekeeping the work.
Partnered with Experimentation to put design hypotheses into live A/B tests — then let the data, not opinion, decide what shipped.
Pushed SplitPay awareness into support landers, promo cards, and help articles — because a great flow nobody can find isn't a great flow.
SplitPay was the first project where I owned both the design direction and the people carrying it. My job was to set the frame, win the approvals, and then make sure two designers could move fast without the experience fragmenting.
Team: 2 designers · Partners: Product, Business, Engineering, Content & Legal, CX, Experimentation, Care, App
Mapped every flow and edge-case state myself — enrollment, payer assignment, mid-cycle changes, cancellation — so each designer inherited a scoped problem, not a blank page.
Presented concept, flows, and trade-offs to Business, Product, and senior leadership partners, and secured approval before the team built against it.
Held daily working sessions with the designers, plus peer reviews and Design Standards reviews — raising the bar through critique, not by redrawing their work.
Gave the design green light at go / no-go checkpoints and held the line on quality at handoff, so what shipped matched what was approved.
SplitPay shipped the way good infrastructure products do: a careful throttled launch, aggressive measurement, and relentless simplification of what we learned was getting in people's way.
Concept prototype, content requirements, and scenario workshops mapping every association state — new customer, mid-cycle change, cancelled primary payer, reactivation.
C.360Assign payers, get an SMS, pay your share through the unauthenticated FastPay flow. Tens of thousands of customers engaged; 32% completed enrollment.
C.045Unlocked an addressable base in the tens of millions. Enrollment completion climbed to 37%, with an 87% payment success rate.
C.015For 2-line accounts, choosing a designated payer is meaningless. We removed the step. Task accomplishment lifted +20.45% at 99.99% confidence.
C.135Expansion to fiber, converged accounts and wearables, with push notifications, stored payment methods, and a view showing only your share.
C.140Months from kickoff · partial since last box · C. = heading · ! to !!! = danger
SplitPay had no reference pattern to borrow. Every screen had to do double duty: teach an unfamiliar idea and survive the messy reality of a billing system. Here is the path a customer actually walks.
The best enrollment flow is worthless if nobody trips over it. I gave SplitPay two homes on the billing hub: a conditional “Did you know? You can Split your bill” banner, and a permanent Split the bill row inside Manage Settings ending in Signup.
The rule behind it was the real design decision: the banner only renders if the account is eligible, and disappears for good once a split group exists — so the page never nags someone who already said yes.
Why this screen carries the business case: enrollment fell 40% when an awareness card was removed from this page, and a later placement test lifted enrollment +105%.
This is the enrolled state on mobile, and it shows where SplitPay ends up once it stops being a new feature and becomes part of how someone runs their account.
Open the balance card's overflow menu and Check SplitPay status sits first — above Set-up AutoPay and Manage payment method. Winning that slot mattered more than any banner: it puts SplitPay alongside the two controls people already trust and return to.
On a phone the hub becomes one long scroll, so the deeper Manage Settings entry sits far below the fold. A persistent shortcut at the top is the difference between a feature people manage and one they forget they turned on.
Nobody had ever split a carrier bill before, so opening with a form would have meant asking for commitment before comprehension. The flow starts with “Say hello to the new experience” — a short explainer video with a clearly available Skip.
Giving Skip equal prominence was deliberate. Education should be available, not mandatory — forcing a tutorial on a confident user is just one more step to abandon.
This is the decision screen — choosing who pays. Above the list sits a value panel, “Why should you split billing with other users?”, spelling out the payoff: automatic monthly reminders, payers settling their own share through FastPay, and less hassle for everyone.
Below that, the task stays deliberately plain: Who is the primary account in this list? and a simple radio list. Reassurance and task are kept separate so the decision never feels cluttered.
Stripped of the education panel, the task is just a question and a list. That clarity is what made the project's biggest win possible: for two-line accounts, asking someone to designate a payer is meaningless — there is only one other person.
We removed the step and tested it. Task accomplishment lifted +20.45% and progression to confirmation +22.86%, both at 99.99% confidence. After launch, completion for two-line accounts more than doubled — from 33% to 74%.
The lesson I keep: the highest-performing thing I designed on SplitPay was a screen I took away.
The review step is where an abstract idea becomes real money. Every payer's share is itemized against Your part and the Current balance, with the full PDF bill one click away. Nobody agrees to split a bill they can't see.
Two consent decisions live here rather than buried in settings. “Set it and forget it!” explains exactly when notifications go out — every month, three days before the due date — and Send payment reminder now lets the account holder kick off the first request immediately.
Because SplitPay texts other people on the customer's behalf, consent had to be explicit and legible. That's a content and legal partnership as much as a design one.
The account holder's core anxiety is simple: did everyone actually pay? Payment status answers it in one scan. Outstanding balances sit in neutral dark type; settled shares turn green and read “Paid” — so the exceptions are what your eye lands on, not the successes.
The primary account holder is badged in the list rather than pulled out of it, keeping one consistent ranking while still signalling who owns the account.
A persistent Pay total outstanding balance button anchors the page — a deliberate escape hatch, because the real-world fallback was always “I'll just pay it myself.”
Enrollment is the easy part. People change roommates, swap payers, and leave plans — so Split Pay settings had to make every one of those moves safe and obvious.
Pay for someone handles the most human edge case in the product: a parent covering a kid's line, or one person picking up a share this month. The popover keeps it to a single question — “Who do you want to pay for?”
Two details do the heavy lifting on trust. “All changes made will take effect from the next billing cycle” sets expectations before the change, not after. And the quiet off-ramp lets people leave without hunting, because a payment product earns trust by how cleanly it lets you out.
Screens use placeholder (FPO) content and fictional account data. Additional flows — no-SMS-device handling and the cancellation off-ramp — are available in the full spec on request.
Every chart below is a design decision with a receipt attached.
Opening SplitPay to AutoPay customers nearly tripled engagement (Phase 1A = 1×) and lifted completion 5 points.
Removing one unnecessary step more than doubled completion after launch.
Design hypotheses that earned their place in production.
All three reached 99.9%+ confidence. Placement is a design decision with revenue attached.
Phase 1A payment mix — evidence that designing for wallet-first payments mattered.
Four in ten payers reached for a digital wallet — the argument for keeping FastPay thumb-friendly and login-free.
Adoption compounds — each enrolled account brings multiple lines and multiple payers with it.
accounts enrolled in SplitPay
lines paid through SplitPay — roughly 4.5 lines per enrolled account
payer groups created — lines paying their own share directly
A first-of-its-kind product doesn't fail at the pixel level — it fails in the gaps between teams. Most of my leverage on SplitPay came from being the consistent thread across those gaps: holding the customer's point of view steady while product, business, engineering, and content each optimized for something different.
SplitPay℠ is an industry-first, user-friendly payment solution that lets customers easily and securely split multi-line wireless bills… driving customer satisfaction and business growth while supporting AT&T's strategy of convergence and digital adoption.
— ATS Optimum Team Award, Gold recognition
Phase 2 takes SplitPay beyond wireless: converged and fiber accounts, wearables, and a personalized in-app experience with push notifications, stored payment methods, and a view showing each person only their portion.
The long-term goal is simple — make sharing a bill feel as natural as sharing the plan.