.png)
Verizon
Telecom · US
9% churn reduction in 4 months
Verizon's Project 624 on Google Gemini reduces churn, improves resolution, and gives the CEO a daily AI sentiment digest
Three integrated AI layers: a 24/7 AI assistant in the MyVerizon app, a daily digest for management summarising all previous day's customer interactions, and a Customer Champion system giving agents real-time context, churn risk signals, and next-best-action recommendations.
Friction
Verizon lost 289,000 postpaid customers in Q1 2024 and consistently trailed T-Mobile and AT&T on satisfaction. The CEO published his personal email in the New York Times, a public signal of how serious the customer experience problem had become.
Breakthrough
Google Gemini deployed across three layers simultaneously. App: AI assistant handling billing, plan changes, and device support 24/7. Management: daily AI-generated digest of all customer service interactions with sentiment analysis. Agents: Customer Champion providing live context, churn risk flags, and next-best-action.
Impact
9% churn reduction in pilot markets within 4 months. 17% higher first-contact resolution. 22% shorter average problem resolution time. 70%+ of store visits resolved in a single visit. 23% shorter average handling time per interaction. CEO receives daily complete sentiment overview of all calls.
Unlock the full analysis with breakthrough, impact, what made it smart and its technical approach below!
Problem
Verizon's customer experience problem was simultaneously a product problem, a service problem, and a strategic problem. Fixing one layer without the others would not move the metrics that mattered.
What made it smart
The three-layer architecture is what distinguishes Project 624 from a standard chatbot deployment. The management digest closes the feedback loop at the top of the organisation. The CEO knows by 8 AM every day exactly what customers are saying. That organisational visibility drives accountability in a way that quarterly NPS surveys cannot.
Technical approach
Google Gemini customised model deployed in the MyVerizon app for self-service. Separate AI pipeline transcribes and summarises all customer service interactions daily, generating sentiment analysis and theme identification for leadership. Customer Champion integrates churn prediction scores, customer history, and live inventory into agent desktops in real time.
Strategic lesson
AI-driven customer experience only improves durably when the insight loop reaches leadership. When the CEO sees yesterday's customer sentiment this morning, every manager knows that customer experience is a first-order priority.
Reflection question
How long does it take for a systematic customer experience problem to reach your leadership team, and what would change if that delay was reduced to 24 hours?