We were wrong about the thing we were most sure of.
AuximOS is a small company in Kozhikode. This is the argument we had with ourselves for eighteen months, and how it ended.
The original plan was the obvious one: multilingual chatbots for patients. WhatsApp enquiries, voice calls, appointment booking. We built it.
Then we spent six months in front of hospital owners, IT directors and operations heads across Kerala, and heard the same sentence so many times it started to sound like policy:
“Patients don't want AI. Patients want humans.”
So we believed them, and we turned the whole company inward. We rebuilt the same infrastructure — multilingual voice, tool calling across HIS and EMR, workflow orchestration — to serve staff instead of patients. Doctor documentation. TPA claim rejections. Bed and theatre scheduling. It worked, and the problems were real and large.
What changed our minds was sitting in a hospital in Kozhikode at nine at night, watching a phone ring at an empty reception desk. It rang out. Then it rang again, ten minutes later, and rang out again.
Nobody in that building would have described that as a problem, because nobody was measuring it. And the objection we had accepted for a year and a half suddenly sounded different. Patients do not want a robot instead of a human. They want an answer instead of a ringtone. Those are not the same objection, and we had spent eighteen months solving the wrong one.
The correction was not to go back. It was to stop choosing. The agent that answers at nine at night and the agent that clears the claim six weeks later are the same system holding the same record of the same patient — and that shared record is worth more than either half is alone. Nobody in this market was selling both, because the front-office vendors and the back-office vendors come from different companies.
So that is what AuximOS is now. Seven agents on the front desk, three inside the building, one memory across all of them.
We are early and we would rather say so. We are in design partnership with a small number of providers, building slowly with people who will tell us the truth. If you are evaluating us, we will tell you exactly what works today, what is still being built, and what we do not do at all.
Six things we will not trade away.
Written down because principles that live only in someone's head get negotiated away in the third quarter.
Both halves, or it is not a system
An agent that books the appointment but cannot see the claim is a chatbot with better manners. The value is in the shared record, so we refuse to sell half of it.
The human is never removed, only relieved
A caller can reach a person at any point. A clinician signs every document. A biller approves every claim. Agents do the waiting, the chasing and the paperwork — not the judgement.
Mid-market providers are the underserved ones
Large chains have IT departments of two hundred. A 180-bed hospital or a four-chair practice has an owner who does IT on Sundays. They need this more and are sold it least.
Multilingual is architecture, not translation
Malayalam, Arabic and English are first-class here, alongside the major Indian languages. Patients code-switch mid-sentence and so do our agents. Bolting translation onto an English system does not survive contact with a real caller.
We will not fake proof
We are pre-launch. There are no customer logos on this site, no press badges, no invented uptime figures, no testimonials. Every transcript here is labelled a demonstration. When we have real numbers you will know, because they will have a customer's name on them.
Built in Kerala, for the corridor it sits on
Kerala and the Gulf are one healthcare market serving the same families across two geographies. We built for that corridor first because we live in it, and because nobody building in San Francisco is going to.
Where this is meant to end up.
Indian healthcare software has been shaped by two things: hospital management systems designed in the 2000s, and chatbot vendors selling the same product to banks, telcos and clinics alike. Neither understands a Thursday morning OPD.
Our goal over the next three years is to be the agent layer for a thousand mid-market providers across India and the GCC. Not the largest hospitals — the middle: the hundred-to-five-hundred bed multi-specialty centres, the polyclinics, and the dental practices that between them deliver most of the country's care and have been waiting a decade for software that respects how they actually run.
If that is a problem you also want to work on, we would like to hear from you.
Come and argue with us.
Whether you run a hospital, advise one, work in Indian healthcare AI, or just want to compare notes — we are easy to reach and we like being challenged on this.
We are pre-launch and taking a small number of design partners in India and the GCC. If that is not you, say so on the call and we will not waste your time.