Headline

AI PM @ CNH | Agentic AI for Insurance & Healthcare | Ex-Paytm, Flipkart | 1x Founder | XLRI

About

I'm the AI PM at CNH, building agentic systems for insurance and healthcare, which means most of the work isn't picking the model. It's the small decisions about when the agent acts on its own, when it hands off to a human, and how it shows its work to someone who has to defend that decision later. Bcz in insurance and healthcare, the cost of being wrong is higher than the cost of being slow, which changes every product decision. Before this, a year at Paytm in loyalty and rewards. The problem was making rewards drive real behavior change, not feel-good clicks. At India scale, that's where you learn that a 1% lift is a real number, and the smallest UX decisions compound into real economics. Before Paytm, Flipkart on the seller logistics side. The problem was that what sellers told us they needed and what was actually breaking their day were two different things. Most of the job was finding the real problem under the stated one. In between, I built aiondemand.in. It lets you switch between Claude, ChatGPT, and Gemini mid-task without losing your context, back when models were half as capable and people actually needed that. Had few paying users for a few months, then ran out of funds, took a job to keep building on the side, and learned I couldn't do both. By the time I figured that out, models had gotten good enough that the problem quietly disappeared. The problem expired before I could scale. 😓 I write here a few times a week about agentic systems, product thinking, and what breaks when you try to ship AI inside a real org. Happy to exchange notes with anyone building in the same space.

Location

Bengaluru

AI stack

antigravity

claude code

claude

codex

gemini

langgraph

Work experience

AI Product Manager at Connect and Heal - CNH Care, Jul 2025 - Present · 1 yr 2 mos, Bangalore Urban, Karnataka, India, At CNH, I work on one core problem in healthcare: insurance claims take too long and break too often. A single claim passes through multiple people — ops, policy teams, doctors — and every handoff adds delay. I’m helping replace parts of this manual flow with AI so claims move faster and more consistently. We started with document understanding (the biggest bottleneck today), and are building towards AI-driven policy checks and medical validation. The goal is simple: reduce claim turnaround time from weeks to days to hours.

Founder at AIOnDemand.in, Apr 2025 - Sep 2025 · 6 mos, Bengaluru, Karnataka, India, I built AIOnDemand after getting frustrated with paying for multiple AI subscriptions and barely using them. Instead of monthly plans, the idea was simple: pay only for the time you actually use AI tools. I built a pay-as-you-use platform giving shared access to tools like ChatGPT, Claude and Perplexity. I did everything end-to-end — product, pricing, tech, and distribution. The product reached ~300 users and paying customers — and then I ran out of money. That experience taught me a hard truth: building a product is not the same as building distribution, and early traction doesn’t guarantee a sustainable business.

Product Manager at Spectra Global, Jun 2024 - Jun 2025 · 1 yr 1 mo, Bengaluru, Karnataka, India, Worked with founders and leadership to fix products that weren’t growing. Helped redesign ad formats and user flows on a digital media platform to improve reach and ad demand. Simplified an internal HR product by fixing confusing workflows so employees actually used it. Also worked on fintech dashboards to help teams spot fraud and compliance issues faster.

Product Manager at Paytm, Jun 2023 - May 2024 · 1 yr, Noida, Uttar Pradesh, India, At Paytm, I worked on fixing moments where users were confused, stuck, or dropped off in the app. I improved decision and exit flows that helped users understand what to do next, reducing bounce rates by ~30% and adding ~₹40L in annual revenue. I also worked closely with customer support to identify repeat issues in chatbot and help flows, cutting support tickets by ~60%. Along the way, I spoke directly with users, redesigned onboarding and discovery flows, and learned how small product changes can impact millions of users at scale.

Product Management Intern at Flipkart, Apr 2022 - Jun 2022 · 3 mos, Bangalore Urban, Karnataka, India, At Flipkart, I worked on reducing losses in the supply chain by fixing how decisions were made during pickups. I identified where products were getting damaged, missed, or wrongly accepted and redesigned pickup checks across brands and categories. These changes helped prevent losses worth several crores across lifestyle and electronics. I also worked on defining clear metrics, simple workflows, and on-ground checks so ops teams could catch issues early instead of fixing them later.

Education

PGDM (BM), Business Management, XLRI Jamshedpur

Bachelor of Technology - BTech, Computer Software Engineering, Andhra university