Company:
Verizon
Category:
Conversational AI
My Role:
Requirement gathering, research, identifying gaps, wireframing & design conceptualisation, stakeholder review and dev handoff.

Discovery
Kickstarting with a cross-functional workshop

Key takeaways
→
The chatbot is a high-traffic, low-friction entry point - ideal for cross-selling
→
Users often arrive with context - personalisation is both possible and expected
→
Strategic conversation design can unlock business goals like conversion, retention & education
→
The biggest gap was between what users needed to know and what the bot was telling them
Auditing the current flow
//Current M+H digital journey

Chatbot shortcomings
Opportunity framing
Turning friction into design opportunity
From our research synthesis, two core "How Might We" questions framed the entire design direction.
UX Opportunity
How might we guide users through a personalized M+H journey without overwhelming them?
Business opportunity
How might we drive conversions by proactively surfacing M+H bundles during high-intent moments?
Design — Proposed Flow
Redesigning the conversation from the ground up
The new flow was built around three principles: meet users where they are, guide rather than push, and personalise at every step. Here's how the conversation was restructured.
01
Contextual greeting
The bot opens with a personalised greeting based on whether the user is a prospect or existing customer — immediately setting the right tone and reducing cognitive load.
02
Intent recognition
Instead of a wall of options, the bot asks one qualifying question to understand what the user is trying to achieve — support, upgrade, or learn more about bundling.
03
Value education
For users with bundle potential, the bot surfaces a clear, visual explanation of M+H benefits tailored to their current plan — making the offer feel relevant, not random.
04
Guided decision
A streamlined 2-step qualification flow replaces the previous multi-branch maze — reducing decision fatigue and keeping users moving toward conversion.
05
Warm handoff
High-intent users are handed off to a live agent or a structured purchase flow with full context — no re-explaining, no friction, no lost leads.

Design — Visual System
Building on Verizon's design system
All designs were built within the Verizon Design System to ensure consistency across the broader digital ecosystem and speed up developer handoff. Key components extended for this project included conversational UI patterns, bundle comparison cards, and progress indicators for multi-step flows.
Impact
Results that moved the business
The redesigned flow was validated through usability testing and piloted with a segment of Verizon's chatbot traffic, delivering measurable improvements across key conversion metrics.
80%
CSAT Surge - improved overall customer satisfaction scores through a seamless, redesigned flow.
↑ FWA
Increased Fixed Wireless Access conversions by leveraging contextual personalization during key user interactions.
↑ M+H
Boosted Mobile + Home bundle conversions by capturing user intent at critical decision-making moments.
Learnings
What this project taught me
Conversation is a UX surface
Chatbot design isn't just copy — it's information architecture, progressive disclosure, and interaction design all at once. Every word choice is a design decision that affects conversion.
Personalisation unlocks trust
Users responded dramatically better to flows that acknowledged what they already had. A simple context check at the start of a conversation changed the entire emotional tone of the interaction.
Business and user goals
align
What was good for users — clarity, relevance, less friction — turned out to be exactly what drove conversions. Designing for the user first was also designing for the business metric.
Workshops accelerate
alignment
Facilitating a cross-functional workshop at the start saved weeks of misalignment later. Getting product, engineering, and business in the same room early meant everyone was solving the same problem.
































































