Dr. Reza Kulkarni runs clinical ops at Lumen Labs, a diagnostic chain with 36 centres across Maharashtra. After three years of nurses fielding appointment changes at all hours, he replaced the entire after-hours phone tree with AIVA — and watched 12,000 patient calls a month resolve themselves.
Half our patients reply to appointment SMS at 2 AM. AIVA handles every one of them — and our nurses don't have to look at their phones until 9.
The challenge
Diagnostic clinics live and die by appointment scheduling. Show up, draw blood, run tests, deliver results, get paid. Miss an appointment and you've burned a slot — at Lumen's volume, that's hundreds of empty chairs a week.
The problem wasn't the bookings. It was the changes. Patients confirm an 8 AM blood test, then text at 11 PM the night before to reschedule — flu, work, school run, the usual. Until 2024, those texts went to a duty nurse's phone. Three nurses on rotation, sleeping with their phones face-up.
The math was painful. Lumen handles roughly 12,000 inbound patient calls a month — most of them about scheduling. About 4,000 come in outside business hours. At one to two minutes per response, that's 120+ hours of unpaid clinical time per month being burned on rescheduling, all of it after dinner.
Three nurses on after-hours rotation, fielding patient SMS until midnight.
Average response time to a reschedule request: 4 hours. Patients gave up and just no-showed.
No-show rate stuck at 14% — nearly 1 in 7 booked slots burned.
Nurse turnover up 22% YoY. The phone duty was the most cited reason.
The setup
Reza had two requirements before signing off. First, the AI had to handle Marathi natively — most of Lumen's patients in Pune and Nashik don't speak Hindi as their first language. Second, anything medically uncertain had to escalate to a human, instantly.
They piloted Voice + SMS over a long weekend. Tuesday morning the assistant took its first patient call — a woman wanting to reschedule her 7 AM cholesterol test. AIVA picked up on the second ring, in Marathi, and rebooked her for Friday. No human looked at the call. By Friday afternoon, AIVA had handled 340 calls and 1,200 SMS replies.
The escalation rules took the most thought. Reza wanted clear lines: any mention of bleeding, chest pain, fainting, or pregnancy goes to a human within 60 seconds. Anything about test results — confidential. Anything about a paid invoice — also human. Everything else, AIVA could handle.
The integration into Lumen's existing scheduling system (a custom HTTP step) took a single afternoon. Their patient records system was already exposed via internal API — AIVA just needed an auth token. Reza signed off on the rollout on Sunday night. The on-call rotation ended on Monday.
The outcomes
The story I tell new nurses now is — you'll never carry a duty phone here. We give you a stethoscope and you keep your evenings. The recruitment difference has been remarkable.
— Dr. Reza Kulkarni, Chief Medical Officer, Lumen Labs
Cut wait time by 84% without growing the support team.
Launched in 3 markets, 3 languages, 3 minutes.
One number, 40 languages, every customer feels heard.