Scaling Support with IVR
What is the project?
A global redesign of the Rider automated phone system (IVR), transforming a rigid legacy menu into an intent-based conversational experience. I led the transition to a brand-aligned voice that manages complex support flows across 24 countries.
What makes it impactful?
Scale: Flawlessly managed 15,000 calls daily with a 100% flow success rate.
Efficiency: Maximized ticket deflection by resolving issues before human intervention.
Trust: Leveraged empathetic pathing to improve rider satisfaction in high-stakes safety/payment crises.
Governance: Scaled a unified voice and personality system across 24 global markets.
How did I build it?
Intent Architecture: Mapped thousands of user intents into logical decision trees and "happy paths".
Voice & Personality: Defined the prosody and tone guidelines to move the system from "robotic" to human-centered.
Conversational Systems: Automating Rider-Customer Support (IVR)
The Challenge: Friction in Human-to-Human Contact
Our legacy support model relied on riders manually calling customers, a process fraught with multiple systemic flaws:
Language Barriers:Â Communication often broke down due to language differences between riders and customers.
Economic Friction:Â Riders faced high personal costs, especially for international calls.
Lack of Governance:Â The organization had zero control or visibility into whether calls actually happened or how long they lasted.
The Strategy: Intent-Based Automation
I worked together with the product team on Interactive Voice Response (IVR)Â system to resolve these issues at scale. By replacing the manual "Call" button with a structured conversational flow, we moved from unpredictable human contact to a governed automated experience:
Automated Persistence:Â The system now automatically attempts to reach the customer up to 3 times, ensuring reliability without rider effort.
Simplified Decision Trees:Â We empowered customers with clear, conversational choice-they can choose to contact the rider or cancel the order directly within the IVR flow.
The Execution: Scaling Logic and Impact
Implementation was focused on removing friction through precise Natural Language Processing (NLP)Â logic and technical integration with Twilio. I designed the prompts to be short and actionable, ensuring that even complex edge cases felt intuitive.
The Impact: Quantifiable Success
The launch delivered immediate, massive results for both the business and user experience:
High-Volume Scale: Successfully managed 15,000 customers daily in the Spain pilot alone.
Operational Excellence: Achieved a 100% success rate on the automated call delivery via Twilio.
Friction Reduction: Post-GoLive data showed a significant drop in both Rider Contact Rates and Fail Rates, specifically reducing "Unable to find" incidents by over 60%.
Efficiency:Â Maintained a concise average call duration of only 31 seconds, dramatically speeding up time-to-resolution.