T-Mobile's IoT solutions help businesses connect, monitor, and manage devices at scale across the company's 5G and LTE networks. The intake framework that drives every deal — vertical, network, hardware, security, bandwidth, international, resource assignment — used to live in a Salesforce dual-list-picker hellscape. This is the redesign.
Three questions about the customer's environment. The framework derives the rest of the security configuration and shows its work — every derived value can be overridden.
Hero · the redesign itself. Section 5 (Security requirements) rebuilt: stepper across the top with completion state, three meta-questions deriving 8 of 11 fields, ownership and downstream consequences visible in the rail. Compare the legacy screen: two bare dropdowns, a disconnected progress rail, and ten more fields hiding behind "Next."
The original Connectivity Framework intake was a 9-step Salesforce form using dual-list-pickers, multi-select dropdowns nested inside dropdowns, and "View all dependencies" links that opened other Salesforce records. Average completion time on the legacy form was 47 minutes for trained IoT AEs. The actual content — vertical, network, hardware, security — was sound. The interface was making the work impossible.
Layered on top: an internal design system built for general Salesforce CRM use, not for technical configuration, and a change process that ran through two PM organizations and an enablement team. What follows is what shipped within those constraints — and, where it matters, the version I'd ship without them.
The intake isn't an isolated form — it's the artifact that crosses five organizational boundaries. Each phase owner picks up a partially-filled framework and either advances it, kicks it back, or escalates. Designing the framework meant designing the handoffs.
Was: SDR collects loose notes in chat. AE re-derives every requirement.
Was: Framework intake started here, mid-funnel. Too late.
Was: Solution architect had to re-interview customer to fill bandwidth + international.
Was: Hardware + security sections often blocked deals at this stage.
Was: Implementation team starting fresh; 30% rework on average.
Reading this as a user journey is the key move. The intake isn't a form — it's a baton. Phase-by-phase responsibility is what the redesign foregrounded.
Scroll — each section takes the stage, then collapses to its title bar above as the next one arrives.† Real fields from the legacy Salesforce form, redesigned. Click any title to jump. Every section applies the same four moves: semantic grouping, gate questions first, smart defaults with visible provenance, downstream consequences stated inline.
† Representation. The production screens live in T-Mobile's Salesforce org and can't be shown directly; this mock is rebuilt in HTML with the real field names, grouping, and derivation logic from the shipped redesign. Data values are illustrative.
Before the redesign, ownership was implicit and contested. The new framework surfaces it inline on every section header. R responsible · A accountable · C consulted · I informed.
The decision graph below shows the dependency chain the redesigned framework enforces. Choosing Vertical = Supply Chain & Logistics narrows Network to 3 options (from 8), which narrows compatible OEM models to 11 (from 30+), which drops the SIM specifications that won't apply.
Fig 1 — Active path highlighted for an example Supply Chain & Logistics opportunity. Each node is a fielded option; edges are downstream constraints that auto-filter the next step.
Fig 2 — The deal-flow dependency graph the framework lives inside. Intake feeds quote logic and custom pricing directly — not just the next step. The dashed edge is the reopen path: renegotiation at contracting edits pricing state, not documents.
The framework is a baton crossing five organizational boundaries. Once that was named explicitly, every interaction decision (chip vs dropdown, conditional show/hide, smart defaults) had a principled answer.
Both are sitting in review. The version on this page is what shipped under enterprise design-system constraints. The unconstrained version would have launched 4 months earlier, with 30% better completion times.
The same system as a 7-beat scroll-driven story: AI lead scoring → account hub.
Walk the system → 02 / 2022–2024Parent case study. Discovery → design system → AI tooling, framed as one program.
Read parent → 03 / 2018–2019Dual-path checkout across two brands from one spec.
Read case →