Back to cases Case 04 · Accenture / Vivo · 2018 — 2019

Automating
Technical Support.

How I reduced connection issue resolution time from 20 minutes to 2 minutes with automated in-app diagnosis.

Client
Vivo (Telefônica)
Period
2018 — 2019
My role
Service Designer
Impact
90% less time
dx
The challenge

Resolve without calling support.

Vivo is Brazil's largest telecom operator. The app served hundreds of thousands of sessions, but when a connection failed, the natural path was to call support. Each case took between 15 and 20 minutes: problem description, manual diagnosis, resolution.

The challenge: build an in-app flow that automatically identified the failure type and resolved it without human intervention. Less time for the customer, less load on operations.

My role

Discovery, blueprint and rapid design.

01
Discovery
Interviews with infrastructure managers and tech leads. Benchmarks. User research on behavior during connection failures.
02
Service Blueprint
End-to-end flow mapping: app, backend, operations. Co-creation with PMs and engineering to align technical capacity and experience.
03
Design
Prototypes and remote usability testing. Contextual CTAs validated with users before handoff.
04
Design System
Mapping the visual chaos across Vivo apps. Evidence that justified the creation of a dedicated team and a full rebrand.
The solution

Four steps, no call center.

Research revealed that the main frustrator wasn't the wait time: it was not knowing what was wrong. The system needed to name the problem before proposing the solution.

Automated journey
01Diagnosis
02Detection
03Contextual CTA
04Confirmation

"Resolving without calling became the standard. That means giving time back to people."

Impact

Less time, more control.

90%
reduction in handling time (20 min → 2 min).
DS
evidence that unlocked Vivo's unified rebrand and Design System.
Resolution time
Before 20 min
After 2 min
Service Design Blueprint Telecom Design System Usability Figma
Learnings

What I take from this case.

· Automated diagnosis requires the system to interpret context, not just technical state. Designing the logic matters as much as designing the screen.

· Contextual CTAs only work when backed by real behavioral data. "Advance promotion" wasn't a generic option: it was the right action at the right moment.

· Design System: documenting the problem is the first step to justifying the structural solution. The visual chaos was visible for years; naming and systematizing it created the momentum to change it.

Next case · 05

AI that recognizes human emotions