Launching in Tamil Nadu: our first 100 doctors

Larinova is launching its pilot program in Tamil Nadu, starting with 100 doctors across Chennai, Coimbatore, and Madurai. We didn't pick Tamil Nadu randomly. It's where code-mixing is hardest, our Sarvam-based STT is strongest, and we can physically visit every pilot clinic within a day's drive.

Why Tamil Nadu first

Tamil Nadu has one of the highest doctor-to-patient ratios in India, with a massive private clinic network. Chennai alone has over 12,000 registered private practitioners. These doctors are tech-comfortable - most already use WhatsApp for patient communication and digital payment apps for billing. But their clinical documentation is still manual. Paper prescriptions, handwritten notes, or at best, typing into basic templates on a laptop between patients. The state also has a strong tradition of Tamil-medium medical education and practice. Doctors here don't just code-mix occasionally - Tamil is the primary language of consultation for the vast majority of patients. If our Tamil+English transcription works here, in the most linguistically demanding environment, it works everywhere.

What pilot doctors told us

We ran a closed alpha with 12 doctors over four weeks. General practitioners, pediatricians, an orthopedic surgeon, and two dermatologists. The feedback reshaped our product in ways we didn't expect.

• Speed over accuracy: Doctors cared more about getting a 90% accurate note in 30 seconds than a 99% accurate note in 2 minutes. They can scan and fix a small error instantly. Waiting kills the workflow.
• Prescription is king: We initially focused on SOAP notes as the primary output. Doctors told us the prescription is what they need first - it's what the patient takes to the pharmacy immediately. SOAP notes are important but secondary in the consultation flow.
• Phone, not tablet: Every pilot doctor used Larinova on their phone, not a tablet or laptop. The device sits on the desk during consultation. They want to tap once to start, once to stop, and see the result. Anything more complex and they'll revert to pen and paper.
• Privacy anxiety is real: Two doctors initially refused to pilot because they were worried about patient recordings being stored or shared. We had to walk them through the architecture - real-time processing, no persistent audio storage - before they'd agree. Trust documentation needs to be front and center.

The 100-doctor pilot structure

Our pilot is structured in three tiers. The first 30 doctors get white-glove onboarding - we visit their clinic, set up the app, observe two consultations, and tune the system for their specific speech patterns and specialty vocabulary. The next 40 doctors get remote onboarding with a dedicated WhatsApp support line and weekly check-in calls. The final 30 doctors get self-serve onboarding with in-app guidance only. This tiered approach lets us measure how much hand-holding the product actually needs. If the self-serve cohort retains as well as the white-glove cohort, we know the product is ready to scale without a field team.

We're not trying to prove that AI medical scribing works. Dragon Medical and Nuance proved that a decade ago in the US. We're proving that it works in Tamil, in a 30-patient-per-day clinic, on a phone, for ₹999 a month.

What's next

After Tamil Nadu, we're expanding to Karnataka (Kannada+English) and Andhra Pradesh (Telugu+English) in Q3 2026. Hindi+English support is already built and will launch alongside a Delhi NCR pilot in Q4. Our roadmap is language-first, not feature-first. Every new state means a new language pair, new medical vocabulary, new regional drug brands, and new consultation patterns. We're not chasing language count. Five languages done right beats twenty done poorly. The first 100 doctors in Tamil Nadu will shape everything that follows. If you're a doctor in Chennai, Coimbatore, or Madurai and want to join the pilot, the waitlist is open.