The Indian hospital pilot that actually converts

Estimated reading time: 5 minutes

Part of Selling AI to Indian Hospitals, a field guide.

In most software sales the pilot is a step before the sale. In Indian healthcare it is the sale. A hospital rarely signs a real contract off a demo and a slide deck. It signs after a pilot proved, on its own patients and its own floor, that the thing works. Which means the free, open-ended, “let’s just try it and see” pilot is where more deals die than anywhere else. It feels like progress. It is usually a slow no.

At BeyondChats we run AI that talks to patients, so almost every deal passes through a pilot. Here is what separates a pilot that converts into a paid contract from one that quietly runs forever and then dies.

A converting pilot has three things, in one line each:

1. One agreed success metric: a single number everyone accepts as the verdict.
2. A fixed end date: a day on the calendar when you decide, not “let’s keep going.”
3. A named owner inside the hospital: one person who has to report the result upward.

1. One agreed success metric, decided before you start

The single most common reason a pilot fails to convert is that nobody agreed, up front, what “success” would look like. Without that number, the end of the pilot becomes an opinion contest, and the hospital’s default opinion is to do nothing.

Pick one metric that the hospital already cares about and that your product plausibly moves: after-hours enquiries captured, the no-show rate on booked appointments, first-response time to a patient message, leads that turned into consultations. Write it down. Agree the baseline: what the number is today, before you arrive, because a pilot with no baseline can’t prove anything. Then agree the threshold: “if we get X, this worked.”

The anti-pattern is two metrics, or five. A pilot measured on everything is measured on nothing, and whatever number looks weakest at the end becomes the reason to delay. One metric, one baseline, one threshold.

2. A fixed end date

“Free forever” is not generous; it is fatal. A pilot with no end date has no moment of decision, so the decision never comes. The product becomes a free utility the hospital quietly depends on and never pays for, and your team spends its life supporting an account that will never close.

Set a real end date: four to eight weeks is usually enough to move one metric in healthcare, and put a review meeting on the calendar for that date at the start, with the decision-maker invited. The end date is what converts “we’re still evaluating” into “we said we’d decide today.” Time-box it, and protect the box: scope creep (“can it also do this other thing?”) is how a four-week pilot becomes a four-month one. New requests go on the post-contract roadmap, not into the pilot.

3. A named owner who reports the result upward

A pilot with no internal owner is a pilot no one at the hospital is accountable for. You need one named person on the hospital side, not “the IT team” but a person, who agreed to run the pilot and, crucially, who has to report the outcome to whoever signs. That person is your champion. Their job is on the line for the result, which means they will chase the data you both need.

If you can’t find that person, that itself is the signal: the hospital is curious, not committed, and the pilot will drift. The named owner is what connects the floor (where the metric moves) to the office (where the contract gets signed). Without them, a great result dies in an inbox.

Agree the conversion trigger before you begin

Here is the move that ties all three together: agree, in writing, what happens if the pilot succeeds, before the pilot starts. “If we hit X by , we sign a contract for Y at ₹Z.” Not a handshake at the end; a pre-agreed trigger at the beginning.

This is uncomfortable to ask for, and that is exactly why it works. A hospital willing to agree the trigger is a real buyer. A hospital that refuses to commit to any outcome, even a successful one, is telling you the pilot is theatre. Better to learn that in week zero than in month four.

The pilot anti-patterns, in one place

• No metric: “let’s just see how it goes.” Nothing to point at when it’s time to sign.
• Too many metrics: the weakest one becomes the excuse to delay.
• No end date: the free pilot the hospital never pays for and never ends.
• No exec sponsor: a great result with no one to carry it upward.
• Scope creep: new asks stretch the pilot past the point of decision.
• No conversion trigger: success was never linked to a signature.

None of this makes a pilot harder to sell. It makes it easier. A hospital that agrees to a metric, a date, an owner and a trigger has already, in effect, decided to buy if you deliver. Your job in the pilot is then simple: deliver.

This is part of what I’m writing about here: selling and building AI for Indian healthcare, in public. If that’s your world too, here’s why I started writing, and the monthly letter below is where the numbers and the messier lessons go.

Next in the field guide: once a pilot converts, someone still has to sign: who actually signs? And the questions that come up in every pilot are the five DPDP data questions. Or see all the guides. For the operations side of that success metric, see where patient flow actually breaks in an Indian hospital.