Curiosity and discipline, engineered into a career.
Hari Kishan is Director of Cloud Engineering at Manulife John Hancock Retirement — an enterprise technology leader specializing in large-scale conversational AI, agentic workflows, and cloud-native customer experience platforms.
Building the systems behind trusted enterprise conversations
Hari led the shift from traditional IVR systems to an AI-augmented contact center built on Amazon Connect — integrating RAG-based SSML optimization, sentiment-aware response generation, and real-time agentic decision layers in a highly regulated financial environment.
He has built several startups from scratch and serves on boards, providing direction and best practices on both technology and business fronts. He is a published author and researcher with multiple publications in the GenAI field.
In August 2026, Insights Success Magazine featured him on the cover of its “Inspirational Icons to Watch in 2026” issue.
The path
Director of Cloud Engineering — Manulife John Hancock Retirement
Leading cloud and AI engineering for enterprise customer experience: conversational AI, agentic workflows, and cloud-native platforms at financial-services scale.
IVR → Agentic Voice AI Transformation
Reimagined a legacy Avaya IVR into an agentic, self-optimizing platform on Amazon Connect with RAG pipelines — SSML models adjusting tone and phrasing in real time, reduced AHT, and an orchestration layer updated in minutes, not weeks.
Startup Builder & Board Advisor
Built multiple startups from the ground up; advises boards on technology direction, coverage, and best practices across the tech and business front.
Published Researcher & Author
Five published papers on agentic AI, voice agents, graph query engines, and planning robustness. Author of The Thinking Voice: Architecting Agentic Voice AI for IVR and Contact Centers.
Business Intelligence → Python → Enterprise Architecture
A journey from BI development through Python engineering and startup leadership to large-scale enterprise AI transformation.
What drives the work
Architecture First
Models are components, not strategy. Specialized systems with defined responsibilities outperform one giant model trying to do everything.
Trust & Governance
Security, privacy, explainability, and human oversight are design requirements — the foundation of customer trust in voice AI.
Human-Centered AI
AI should enhance employees, not remove human judgment. People validate important outputs and enforce governance controls.
Let’s talk about what’s next.
Advisory, speaking engagements, research collaboration, or media — one conversation to start.