AI Projects — Mayukhee Das

Inbound Voice Automation

24/7 AI voice agent for revenue capture and continuity

Voice AI • CRM Integration • Python NLP • Cloud-Native Telephony

Impact

Reduced TCO by 20% and reclaimed $1.5M+ in operational capacity by eliminating missed after-hours, overflow, and outage-related calls.

Solution

Built a resilient AI voice agent that answered inbound calls, scheduled appointments in the CRM, and escalated emergency scenarios immediately.

Technical Depth

Cloud-native telephony, CRM/ticketing webhooks, Python NLP intent classification, low-latency scheduling logic, CI/CD, and state synchronization for high availability.

AI-to-Human Escalation Orchestration

Warm handoff orchestration between AI and live agents

Escalation Design • Microservices • Sentiment Routing • Agent Desktop

Impact

Improved booking and resolution performance to 85% by reducing friction during AI-to-human transfers and preserving customer context.

Solution

Engineered an escalation layer that detected stalled conversations, complex edge cases, or high-empathy moments and transferred customers without context loss.

Technical Depth

Decoupled microservices, custom session-state management, sentiment-aware routing, and real-time metadata packaging for the agent desktop.

Autonomous Revenue Re-Engagement

Event-driven outbound engine for dormant revenue recovery

Outbound Automation • CRM Pipelines • Customer Segmentation • Revenue Ops

Impact

Converted unbooked leads and idle technician capacity into revenue opportunities without adding scheduling headcount.

Solution

Built a follow-up engine that monitored dispatch-board availability and triggered SMS, email, or voice campaigns for unsold estimates and lapsed agreements.

Technical Depth

CRM-to-dispatch data pipelines, scheduled Python microservices, conditional API webhooks, customer segmentation, and channel-preference logic.

Real-Time Performance Analytics and Call Scoring

100% call coverage for QA, compliance, and coaching

NLP • Streaming Transcription • QA Automation • Tableau

Impact

Expanded QA visibility from limited manual sampling to 100% interaction scoring, improving script adherence, coaching precision, and sales transparency.

Solution

Led an analytics pipeline that transcribed, ingested, and scored every agent call against rubrics for sales compliance, objection handling, and process adherence.

Technical Depth

Streaming transcription APIs, NLP classification, SQL transformations, deterministic scoring rules, and Tableau dashboards for executive visibility.

Tier-1 Automation Engine

Conversational AI for high-volume support deflection

Cognigy.AI • NLU • ITSM Integration • Workflow Automation

Impact

Achieved 60% Tier-1 ticket automation, 80% AI-assisted FCR, 92% intent recognition accuracy, and 50% lower human-in-the-loop dependency.

Solution

Architected a proactive automation ecosystem that resolved repetitive Tier-1 requests through conversational flows, auto-triage, and backend workflow execution.

Technical Depth

Cognigy.AI, Cognigy Voice Gateway, native NLU models, REST microservices, ITSM/CRM integration, and secure workflow execution patterns.

Context-Driven IVR and Agentic AI

Dynamic intent routing with Omilia, Genesys, and NICE

Omilia • Genesys • NICE • Agentic Routing • IVR Modernisation

Impact

Reduced call misrouting by 40%, lowered AHT drivers, and improved upsell conversion by giving agents full caller context before answer.

Solution

Replaced rigid IVR decision trees with AI-driven pathfinding that interpreted caller intent, routed dynamically, and supported warm handoffs.

Technical Depth

Omilia agentic pathfinder integrated into Genesys/NICE, data-ingestion pipelines, context-transfer APIs, and unified desktop metadata delivery.

← Back to Portfolio