Open to senior AI product roles & advisory

Enterprise AI that survives contact with production.

I'm Prateek Kunwar — an AI Product Manager with eight years bridging complex AI engineering and business strategy. I build Generative AI, NLP and conversational systems inside regulated enterprises, where a wrong answer is a compliance finding rather than a bad demo.

Email me WhatsApp Download CV (PDF)
  • 35% Fewer support tickets, globally
  • 200% Growth in enterprise AI adoption
  • 8+ Years shipping AI in production
Prateek Kunwar, AI Product Manager based in Bengaluru
Bengaluru, India

Eight years across

Deloitte HSBC Accenture ISB Hyderabad IIM Indore

01Selected work

Three programmes, and what actually made them work.

Most AI initiatives die somewhere between the model and the mandate. These are three that didn't — and the decisions that were the difference.

01 Deloitte · Global Service Desk

An AI service desk that removed a third of global tickets

Challenge
Tier-1 support volume was growing faster than the headcount available to absorb it. Earlier automation attempts had stalled, because they went after the most visible ticket types rather than the most tractable ones.
Approach
I led the end-to-end product lifecycle for an AI-powered ServiceNow Virtual Agent, owning the backlog and writing the user stories myself. We sequenced intents by volume-to-complexity ratio rather than by stakeholder noise, and stood up a defect-triage framework that ranked bugs by business impact instead of by report date.
Outcome
  • 35%Reduction in global ticket volume
  • 30%Faster mean time to repair
02 Deloitte · Tier-1 Indian Bank

A GenAI contact centre built inside a bank's compliance envelope

Challenge
A Tier-1 Indian bank wanted generative AI across both inbound service and outbound sales. In retail banking, a wrong answer isn't an embarrassing screenshot — it's a regulatory finding, so the usual "ship it and iterate" route was closed.
Approach
I defined and owned the product roadmap, translating sophisticated banking requirements into scalable AI models. Governance was treated as an architectural constraint from day one — data handling, grounding and evaluation designed in rather than retrofitted in the fortnight before launch.
Outcome
Deployed into production across inbound and outbound sales campaigns, meeting the bank's AI governance and compliance standards — and reducing MTTR across the wider global banking application estate by 30%.
03 Deloitte · Sales & Revenue

The NLP engine that finally got used

Challenge
Sales optimisation models already existed and performed well in evaluation. Adoption was flat. A model nobody opens returns nothing, however good its benchmarks — so the problem was never really a modelling problem.
Approach
I drove enterprise-wide scaling on the back of predictive analytics, shipped a generative text model into production for sales teams, and built a C-suite ROI reporting framework on SAP analytics and BI dashboards — so the people funding the programme could see what it returned without asking anyone.
Outcome
  • 200%Growth in user adoption
  • 40%Sustained lead conversion rate
  • 25%Better client targeting accuracy

02By the numbers

Eight years, measured.

Every figure below belongs to a system that reached production and real users — not a prototype, and not a slide.

  • 35% Reduction in global support ticket volume Deloitte · AI service desk
  • 200% Growth in enterprise AI user adoption Deloitte · NLP sales engines
  • 15K+ Monthly data requests automated Deloitte · AI-powered SAP interface
  • 30% Faster incident resolution across banking apps Deloitte · Global banking estate
  • $500K Annual cost savings from process optimisation HSBC · Business consulting
  • 10K+ Monthly chatbot conversations at 90% satisfaction Accenture · RASA & Dialogflow

03Approach

Three convictions I keep being proved right about.

01

Start from the mandate, not the model.

Most failed AI programmes begin with a capability looking for a problem. I start from the regulatory or commercial obligation the business already carries, then work backwards to the smallest AI capability that discharges it. It is a less exciting first meeting and a considerably better second year.

02

Governance is a product requirement, not a gate at the end.

In banking and financial services, AI governance shapes architecture, data handling and evaluation — it is not a compliance review bolted on before launch. Teams that treat it as a final checkpoint discover in month nine that the design cannot satisfy it. Building it in from day one is what lets a product actually ship.

03

Adoption is the only honest metric.

A model that nobody opens has a return of zero, whatever its benchmark scores say. I would rather ship a narrower capability that three thousand people use daily than a sophisticated one that impresses a steering committee and then quietly decays. Every number on this page is an adoption or business-outcome figure for that reason.

04Capability

Where I'm genuinely useful.

01 — Product

AI product strategy & lifecycle

Owning the roadmap from discovery to production: backlog, user stories, prioritisation and go-to-market for AI capabilities inside regulated enterprises.

Product Strategy Backlog Prioritisation User Stories Go-to-Market AI Governance Agile Delivery
02 — Build

Generative & conversational AI

Taking LLM and virtual-agent products through the parts that actually break — grounding, compliance, evaluation and adoption at enterprise scale.

Large Language Models Conversational AI Agentic AI Prompt Engineering NLP RASA Kore.ai
03 — Prove

Data & automation at scale

Leading engineering, data science and UX teams to build the pipelines, dashboards and automation that make AI products measurable and defensible.

Python ETL Pipelines Anomaly Detection Power BI Automation Anywhere ServiceNow JIRA

05Career

From building the chatbots to owning the roadmap.

I started as an engineer shipping conversational AI. That grounding is why the requirements I write tend to be buildable.

Manager, Strategy & Operations

Deloitte
2025 — Present
  • Owns the end-to-end product lifecycle for an AI-powered ServiceNow Virtual Agent automating Tier-1 support.
  • Defined the roadmap for a GenAI contact centre serving a Tier-1 Indian bank under strict AI governance.
  • Scaled NLP sales optimisation engines enterprise-wide to 200% adoption growth.

Deputy Manager, Strategy & Operations

Deloitte
2022 — 2025
  • Shipped a generative text model to production: +25% targeting accuracy, +40% conversions.
  • Engineered an AI-powered SAP query interface automating 15,000+ monthly requests.
  • Directed a multi-million-dollar AI portfolio across concurrent GenAI deployments, under budget.

Senior Analyst, Business Consulting

HSBC
2021 — 2022
  • Eliminated 20% of redundant workflows for $500K in annual savings.
  • Built a productivity dashboard for 500+ users, improving task efficiency 35%.

Senior Software Engineer

Accenture
2018 — 2021
  • Deployed 15+ chatbots on RASA and Dialogflow handling 10K+ monthly conversations.
  • Built an Excel-to-XML automation tool that removed 200 hours of manual work monthly.

Full career detail

06Credentials

Trained where strategy meets engineering.

Indian School of Business, Hyderabad

Advanced Management in Business Analytics

2021 — 2022
IIM Indore

Integrated Programme in Business Analytics

2019 — 2021
KIIT University, Bhubaneswar

Bachelor of Technology

2014 — 2018

Certifications

OpenAI Certified Agentic AI Certification ServiceNow Virtual Agent Kore.ai Advanced Certified Certified RASA Developer Automation Anywhere Advanced RPA Prompt Engineering for Developers Building a Database Agent with GenAI n8n Certified LivePerson Certified Professional

ACE Award

Deloitte. Honoured for exceptional contribution to key projects and outstanding performance in delivering impactful business outcomes.

Tech Star Award

Recognised for innovation in technology — pioneering solutions that drove measurable improvement and strategic value.

07Questions

The short answers.

Who is Prateek Kunwar?

Prateek Kunwar is an AI Product Manager based in Bengaluru, India, with 8+ years of B2B product lifecycle experience. He currently works at Deloitte as Manager, Strategy and Operations, and has previously held roles at HSBC and Accenture. He specialises in taking enterprise Generative AI, NLP and Conversational AI products from strategy through to production deployment in regulated industries.

What does he specialise in?

Prateek Kunwar specialises in AI product management for regulated enterprises. His core areas are Generative AI and large language model products, conversational AI and virtual agents, NLP-driven sales and service optimisation, AI governance and compliance, and intelligent automation. He is known for translating complex regulatory and business mandates into AI capabilities that reach high adoption.

What results has he actually delivered?

He has reduced global support ticket volume by 35% through an AI-powered ServiceNow Virtual Agent, grown adoption of NLP sales optimisation engines by 200% while sustaining a 40% lead conversion rate, cut mean time to repair by 30% across global banking applications, automated more than 15,000 monthly data requests through an AI-powered SAP query interface, and delivered $500,000 in annual cost savings at HSBC.

What is his educational background?

Prateek Kunwar holds an Advanced Management programme in Business Analytics from the Indian School of Business (ISB) Hyderabad, an Integrated Programme in Business Analytics from IIM Indore, and a Bachelor of Technology from KIIT University, Bhubaneswar. He is OpenAI Certified and holds Agentic AI, RASA, Kore.ai, ServiceNow Virtual Agent and Automation Anywhere Advanced RPA certifications.

Is he available for roles or advisory work?

Yes. Prateek Kunwar is open to senior AI product management roles and to enterprise Generative AI advisory engagements. He can be reached by email at prateekkunwar180@gmail.com or on WhatsApp at +91 78548 07551.

Get in touch

Let's build something worth measuring.

Whether you're hiring for a senior AI product role or trying to get a Generative AI programme past the pilot stage — I read every message myself, and I reply.

Emailprateekkunwar180@gmail.com WhatsApp+91 78548 07551 LinkedIn/in/prateek-kunwar