Voox.ai
“These guys have become an extension of my team, they are trustworthy, supportive and rapid when it comes to development or solving problems. An amazing experience as always. Thank you.”
The Problem
Voox.ai was born from a clear insight: call centers generate massive amounts of conversational data, but most of it goes unanalyzed. The client, CEO Alessandro Giordani, had a vision for an AI-powered platform that could surface insights from every call (sentiment, agent performance, compliance flags), but needed an expert team to build the MVP and validate the concept.
The technical challenge was significant: processing audio from hundreds of simultaneous calls, extracting structured insights through deep learning and NLP, and presenting actionable data to contact center managers in real time.
Our Approach
We started with an MVP built on Node.js and React, using MongoDB for flexible data storage. The architecture was designed from day one to be serverless: AWS Transcribe for speech-to-text, Lambda functions for serverless call analysis, and S3 for storing call recordings and data files. This kept costs low during the MVP phase while allowing the platform to scale to enterprise volumes.
The analysis pipeline processes calls through multiple stages: transcription via AWS Transcribe, sentiment scoring, keyword detection (positive language, empathy markers, swears, compliance phrases), and agent performance metrics. Managers get dashboards showing AHT (Average Handle Time), FCR (First Call Resolution), agent performance scores, and compliance rates, with drill-down into individual call transcripts.
We built smart filters that let managers slice data by agent, team, time period, or sentiment category. The chat-like call transcript view makes it easy to jump to specific moments in a conversation without listening to the full recording.
The Results
The MVP validated the concept decisively. Voox.ai now brings agent assistance, automatic speech recognition, and NLP to modern contact centers. The client expanded from MVP to a full product used by multiple BPOs and call centers. The results speak for themselves: 4x increase in bookings, 40% increase in conversion rate, 5% increase in CSAT, and 87% improvement in QA efficiency.
Outcomes
- ◆4x increase in bookings post-MVP
- ◆40% increase in conversion rate
- ◆5% increase in CSAT scores
- ◆87% increase in QA efficiency
- ◆Serverless AWS architecture scales to hundreds of concurrent calls
- ◆Multi-language support for international contact centers