← Back to All Projects
AI / ML

Tele-Marketing Call Audit Automation

Just Dial — Internal ML Product

Designed and deployed LSTM-based deep learning models to automatically audit tele-marketing call quality — replacing a fully manual review process and scaling to 3 lakh+ minutes of audio per month with near-zero human effort.

3L+Minutes Audited / Month
95%+Classification Accuracy
PyTorchLSTMPythonAudio MLFlaskMySQLDocker

The Challenge

Just Dial operates one of India's largest tele-marketing operations, generating hundreds of thousands of minutes of call recordings monthly. Quality audit was entirely manual — a team of reviewers listened to random samples, an inherently slow and unscalable process that covered only a fraction of total calls.

The goal was to build an automated ML pipeline that could process every call, classify quality parameters, detect compliance issues, and flag problematic calls — all without human involvement.

Key Features

🎙️

Audio Feature Extraction

Custom audio preprocessing pipeline extracting MFCCs, spectral features, and temporal patterns from raw call recordings.

🧠

LSTM Classification Models

Sequence-to-label LSTM models trained on labeled call data to classify call quality, compliance, and script adherence.

📊

Automated Reporting

Dashboard showing audit results, flagged calls, trend analysis, and per-agent performance metrics in real time.

🚨

Anomaly Detection

Automatic flagging of calls that deviate significantly from expected patterns — catching compliance violations early.

Technical Approach

01

Data Pipeline

Built an end-to-end pipeline from raw call recordings to processed features, handling audio normalization, segmentation, and feature extraction at scale.

02

Model Training

Trained LSTM models in PyTorch on labeled call data, iterating on architecture and hyperparameters to achieve 95%+ accuracy on quality classification.

03

Production Deployment

Deployed as a containerized service processing call recordings in batches, with results feeding into the audit dashboard and alerting systems.

Outcomes

  • 3L+ minutes audited monthlyfrom a fraction to 100% coverage of all calls
  • Near-zero manual efforteliminated the need for a dedicated manual audit team
  • 95%+ accuracyML models matching or exceeding human reviewer consistency
  • Faster feedback loopsagents get quality scores within hours instead of weeks
  • Compliance enforcementautomated detection of script deviations and violations

Interested in building audio or NLP?

I build deep learning pipelines from data processing to production deployment.

← PreviousReal-Time TME Assist Platform
Next ProjectAgentic E-Care Portal