AI Voice Chat & Self-Buy Flow
WheelsEye is a fleet management platform in India offering telematics solutions like - GPS tracking, advanced Pro features, diesel monitoring and many more. It serves logistics companies and large fleet owners seeking operational visibility, fuel control, and compliance management at scale - shifting manual operations to digital fleet intelligence.
✦ Solely Led
THE GAP
Wheelseye's acquisition was entirely sales assisted. The Self-Buy flow unlocks independent plan selection, checkout and onboarding - a new self-serve growth channel for smaller fleet operators.
💰 Revenue leakage
Order splitting, fake vehicle numbers, and pricing manipulation were hard to catch in a fully human flow.
👨🏻💼 SE as bottleneck
Every order required a sales executive to be present. Scale was capped by headcount, not demand.
🙅🏻♂️ No customer ownership
Fleet owners had zero visibility into their own order, they called the SE to find out what was happening.
📝 Opaque pricing
Customers didn't know if the price was fair. Trust was built on the SE's personality, not the product.
THE CHALLENGE
Sales assited, not Sales dependent
The goal wasn't to remove the sales executive, it was to stop the transaction from depending on them. The SE stays in the loop for context, trust, and attribution. But the customer completes the order themselves.
CORE Decisions
Fleet owners buying GPS devices are not digital first users. They're truck operators used to negotiating face-to-face. Every step had to feel clear, safe, and worth completing without calling someone.
Pricing is transparent from login
No negotiation. No asking. Personalised price shown on first open + progressive disclosure of relevant features.
Pricing is transparent from login
No negotiation. No asking. Personalised price shown on first open.
App executes
Transaction happens in the app. SE assists but never blocks.
Sales has visibility, not control
SE monitors in Thor. Customer owns the order.
DESIGNS
From download to device installation
STATUS
Work in Progress & Learnings
✦ This project is currently in progress.
Design is in active iteration — bulk ordering, drop-off handling, upgrade flows, and the upsell experience are being worked on. Final screens and outcomes will be added as the product ships.
LEARNINGS
Deep product context = faster, better decisions
The Self-Buy flow required understanding pricing logic, plan structures, onboarding sequences, and technical constraints — knowledge built over 9 months. I could move straight into design without re-discovery. That speed and independence is what senior ownership looks like in practice
What's being built next
• Bulk order flow
• Drop-off and resume handling
• Upsell experience design
THE GAP
✦ Mentored + Collaborated
Repeatable queries that didnt need a human agent
30-40% Inbound Calls were informational queries
Fleet operators often knew what they wanted to find, but couldn't navigate to it quickly enough and would call customer support.
Top Informational Query Types
🚛 GPS pricing & feature explanations
💳 Renewal payment queries
🔄 Recharge status / pending queries
📡 Device status / offline questions
💰 Toll & wallet deduction explanation
My Role: Direction & Mentoring
Defined scope, constraints, and UX
principles for the intern to work within
Ran weekly critique sessions on flows and interaction details
Made final calls on ambiguous UX decisions and edge cases
Intern Ownership
UI research and market study
First-pass explorations for empty states and loading patterns
Documentation of chat patterns for developer handoff
RESEARCH
Key Design Decisions
1
Challenge
What should the AI confidently answer vs. gracefully decline, scope needed to be clear upfront
Decision
Defined a strict "answerable" answers which were tied to existing data modules like trips, fuel, alerts.
Rationale
An AI that over promises and fails breaks trust faster than one that's honest about its boundaries from the start.
2
Challenge
Intern was designing screens without a strong understanding of what and how operators actually ask query.
Decision
Created a set of 12 "real operator questions" as design constraints intern had to validate every flow against them
Rationale
Grounding the intern for in real user scenarios, preventing generic chat UX patterns that wouldn't fit a B2B fleet app.
