Generative AI in healthcare: where it works in India today

Transcript

Most hospital AI used to be predictive: scoring a scan or forecasting bed demand.

Generative AI is different.

It writes, talks and summarises.

That makes it good at conversation, drafting and search, and unreliable wherever a confident wrong answer could hurt someone.

Here is where it pays back in Indian hospitals today.

The patient front door: answering logistics on WhatsApp and chat, in several languages, and routing enquiries.

The contact centre: summarising calls and helping agents find answers.

Content and AI search: drafting patient information for clinician review.

And documentation support and internal knowledge, with a human checking the output.

And where it should not go yet.

Diagnosis or treatment advice from a general-purpose model without a clinician in the loop.

Unsupervised medical answers on chat or voice.

Route, do not advise.

Anything that hides the human.

Patients should know when they are talking to software.

And personal health data in tools whose data handling you do not control.

Three things make Indian deployments different.

Language.

Patients switch between English, Hindi and regional languages, often in one message.

Channel.

WhatsApp and the phone carry far more of the conversation than web forms.

And regulation.

The DPDP Act and its Rules shape consent, notice and retention.

A ninety-day way to start.

Weeks one and two.

Pick one problem with a number attached, such as unanswered WhatsApp enquiries after eight at night, and record the baseline.

Weeks three to six.

Build a narrow pilot with a clear handover to humans and reviewed scripts.

Weeks seven to ten.

Run it on real traffic and review conversations every week.

Weeks eleven to thirteen.

Compare with the baseline, then scale, fix or stop, and write down why.

Measure what the business already tracks: enquiries answered on time, appointments booked, calls handled without a callback.

Avoid vanity measures such as messages sent.

The full guide, with sources and checklists, is on gauravphogat.com.

The link is in the description.