From Call Centre
to Self-Serve

A gap that sent
thousands to the phone

Every month, tens of thousands of Telstra customers finish their device contracts, but this milestone lacked digital self-service functionality. The website had no way to let an existing customer buy a new phone without also forcing them to select a brand-new plan. This architecture resulted in duplicate services in the cart, confused users, and a helpdesk fielding thousands of unnecessary calls a year.

The goal was to build a self-serve upgrade journey from scratch, one that allowed existing customers to buy a new device on contract while keeping their current plan, successfully extending their lifecycle by 12 to 36 months.

Broken
by design

The first thing I needed to understand was why this operational gap existed and whether any competitors had solved it. The legacy Telstra.com flow assumed every device buyer was a new customer who needed a plan, leaving a massive percentage of returning users completely un-catered to.

Competitor mapping

I evaluated how Optus, Vodafone, and AT&T handled contract renewals. Optus and Vodafone suffered from the exact same problem we did, forcing plan selection before identifying if the user was new or existing. AT&T was the only competitor handling it differently: they prompted a sign-in immediately after device configuration, allowing returning customers to skip plan selection entirely.

Design Jam

I ran a Design Jam with senior Telstra UX designers. We used a Rose, Thorn, Bud framework on a Miro board to debate scalability, early authentication, and technical feasibility. We landed on three low-fidelity concepts, all built around the same core idea: sign the user in early, check their eligibility immediately, and skip plan selection entirely for existing customers.

Three walls,
three workarounds

Getting tech on board

I presented the concepts to the solution architecture and engineering teams. The major requirement was an early sign-in process built directly into the product pages, a first for Telstra.com. While this flagged technical complexities, it was deemed feasible, especially given its potential to reduce call centre volume post-launch. From that point, the moment a customer signed in on a product page, the system could pull their real-time account data and active services directly into the interface.

Legal and billing

In one scenario, customers upgrading mid-contract owed a lump-sum device buyout fee. Because the backend couldn't calculate this price dynamically, the fee had to be billed separately, creating a strict legal compliance obligation. I collaborated with the legal team to design clear service-update indicators on the product tiles and a dedicated billing disclaimer within the pricing component, ensuring transparency even if the exact figure arrived later. Subsequent user testing validated this approach: all participants were aware of their remaining balance and felt comfortable with the delayed billing structure.

Design system

My first instinct was to use existing design system components to handle the "New vs. Existing Customer" toggle to keep the build simple for the dev team. However, the existing components didn't work for this specific use case; they broke with longer text strings, lacked appropriate visual prominence, and caused stakeholder friction. Following a review with design leadership, I co-designed a custom tab-toggle component alongside our Visual Designer that matched the product page aesthetic and handled the operational need perfectly.

What users
actually told us

I ran two rounds of fully moderated testing with 12 real Telstra customers who managed multi-service accounts and fit our exact target profile. While every participant successfully completed the upgrade task, testing exposed two critical issues we couldn't ignore.

Plan confusion

The product page copy led users to believe they were changing or replacing their core plan rather than keeping it. Because this was a fundamental misread of the entire journey, we went back and completely rewrote the messaging framework for clarity.

The pricing black box

Engineering initially pushed back on surfacing mid-contract buyout costs directly on the product page, citing backend calculation complexities. We tested an alternative version that told users the final price would be provided on a later bill. The user reaction was immediate.

Showing engineering the video clips of 12 participants stating they would abandon the flow provided all the empirical leverage I needed to shift the technical decision. We pushed successfully to get the buyout costs surfaced upfront, eliminating a massive cart abandonment risk.

With the pricing issue resolved, we moved into final refinements. I worked closely with the content and visual design team to tighten the messaging framework, align the UI to the design system, and prepare a thorough handoff for engineering.

Numbers that
actually moved

Following a full production deployment, which included mapping and verifying every technical edge-case scenario alongside the development team, the metrics proved the impact.

What I learned

A 10% reduction in call centre volume in the first month isn't a design win in isolation; it happened because design, tech, legal, and content were all aligned and solving the same problem at the same time. Our pilot structure, relying on daily, direct collaboration between design and engineering, made this cross-functional success possible in a way a traditional waterfall handoff model never could.