Selling AI to Indian Hospitals: a founder’s field guide

Most writing about “selling AI” assumes a US SaaS buyer: a credit card, a two-week trial, a champion who can sign. Almost none of it survives contact with an Indian hospital, where the buyer is a committee, the budget cycle is a year, the law is different, and the pilot is the sale. This is a running field guide to the parts nobody writes down.

I’m Pankaj Baranwal, co-founder and CEO of BeyondChats, where we build AI that talks to patients and prospective patients. I’m writing this in public, partly to help the next founder skip the expensive lessons, and partly because writing it down is how I think. It builds one guide at a time; new pieces publish through the autumn.

Who this is for

• Founders selling any software, not just AI, into Indian hospitals and clinics.
• Product and sales teams trying to turn a hospital pilot into a signed contract.
• Hospital and clinic leaders sizing up an AI vendor. The questions here are the ones worth asking.

The guides

Published

• DPDP, not HIPAA: the 5 data questions Indian hospitals ask AI vendors: why the American compliance pitch fails, and the five patient-data questions that actually decide whether you clear a hospital’s IT gate.
• Who actually signs? Reading the buyer: the committee (superintendent, IT/CISO, finance) versus the owner-doctor versus the tender portal, and how to tell which room you’re in.
• The pilot that actually converts: in Indian healthcare the pilot is the sale. The three things a converting pilot needs are one metric, a fixed end date, and a named owner, plus the trap of the free pilot that never ends.
• Winning government hospital tenders: GeM and the state e-procurement portals, the eligibility clauses that disqualify you before the demo, and why L1 pricing usually decides the winner.
• Integrating with the hospital: HIS/HMIS, WhatsApp and the front desk: the three seams where deals get real: the system of record, the channel patients actually use, and the receptionist whose day you change.
• The objections you’ll hear, and how to answer them: no budget, patients won’t use it, is our data safe, we tried AI before, no time, will it replace staff. What each really means, and why a pilot answers most of them.
• Why deals lapse at renewal, and how to earn year two, the churn nobody plans for: value that never showed up in their numbers, a champion who moved on, the quiet switch-off. How to defend the second year before the renewal meeting.

In the works: the topics I’m drafting next:

• Pricing in ₹, and references and case studies: the rest of the motion once you are in. Both need real numbers, so they come after.

Related reading: why healthcare AI has a trust problem, and what actually earns it. And a growing set of notes on the operations problem the buyer feels: where patient flow breaks, how to cut patient no-shows, and managing the OPD queue.