One system, from first contact to restored sight.
The surgery is five minutes. Everything around it is the system.
Nine stages, from finding a patient to looking after them afterwards. Each shows who does the work, how it runs today, the data it captures, and where an AI agent takes it over.
Real system first. Agents on top.
We do not start with agents. An agent layered on a broken process just breaks faster. We make the journey work — one form, one record, one board — then agents take it over step by step until they are coordinating with each other.
Who does the work: 19 steps run by an AI agent alone, 9 where the agent prepares and a person decides, 13 by fixed software rules, and 4 purely by people — doorstep visits, the surgeon's examination, filing on the government portal, and clinic partnerships. Every clinical decision sits with a person. No agent ever diagnoses.
Foundation
The real working system. One intake form, one patient record, one surgery board, one dashboard.
Automate
Rules and the first agents — camp capture, the qualification call, barrier-matched follow-up.
Agentic
Agents own whole steps. Scheduling, claims assembly and aftercare run without a person driving them.
Self-improving
A supervisor agent coordinates the rest. Changes are tested against real past journeys before release.
Nine stages, end to end
Acquisition through to reporting. Tap any step for what it is, how it runs today, the data it captures, and where the AI takes over.
The screens, stage by stage
One web application for the team, built to work on a desk machine at reception and on a phone in the field. There is no patient app — patients only ever receive a WhatsApp message or a phone call, and the team sees all of it inside the same system.
These are not final designs. They are here to show the shape of the system and what each screen has to do. Layout, wording and visual detail will be worked through with the team.
One centre proves it. The engine is the asset.
The first centre is not the product. The product is a packaged operating engine that turns an empty floor into a working eye hospital — and it gets faster with every centre that runs on it.
The first centre works
Intake digitised, every patient qualified before they travel, the funnel visible stage by stage for the first time.
Documentation holds
Cover recorded at registration, packets assembled and checked, every rejection traced back to the step that caused it.
The next centre is a copy
Multi-centre from day one, so a new site is configuration rather than a rebuild. Weeks, not quarters.
Agents run the network
A supervisor agent coordinates across centres. Every change is tested against real past journeys before it touches a patient.
The part worth pitching
Free eye care in India is delivered on registers and goodwill. Once this engine runs at one centre, it is the thing that makes the next one credible — a proven system rather than an intention. And it transfers: the same machinery fits any free, high-volume, single-procedure service where patients are found in the field and documented for a scheme.