Â
Every day, three different departments on Indian Railways request maintenance blocks on the same tracks. Earlier, this all happened manually : on spreadsheets. No system decided what was actually urgent.
But cut to 2025: Indian Railways launched an official system for this: the Block and Disconnection Management System, BDMS : mandated nationwide by the Railway Board. It digitized how blocks get requested and approved. But we read the official manual ourselves : there's zero AI, zero prioritization logic anywhere in it. Departments still decide manually what gets fixed first. That's the exact gap: and we built Niyojan AI to close it.
Niyojan AI is a layer that sits on top of Indian Railways' existing systems — BDMS and COA. It reads defect records and corridor availability data, scores every defect's real urgency using a trained AI model, then a constraint solver : CP-SAT : builds a conflict-free schedule automatically. And finally, every decision still goes to a human for final approval.
Here are a thousand maintenance defect records, modeled on real railway standards and already scored by our AI. Let's understand how, with one example.
Take this defect : ID 20260656. Niyojan scored it High: 96.86, roughly 97 out of 100. Why?
lets understand this with an example:
Imagine a broken power line hanging directly over the busiest highway in the city. If that same power line were over a quiet village road, we could probably wait a few days to fix it. But this defect sits on the Ambarnath–Badlapur route, which has 100% traffic density. That means trains are running back-to-back, constantly..
The defect is of the department TRD (Traction Distribution), which means it's a problem with the overhead electrical wires that power the trains. Because it is 4 days overdue, the wire is getting weaker and weaker every single time a train passes under it.
If that wire snaps, it doesn't just delay one train. It cuts the power to the entire track, instantly trapping dozens of trains and thousands of people behind it. The AI scores it this high because ignoring an electrical fault on a jam-packed route risks a chain-reaction failure across the whole line.
And here's the result — our AI-generated weekly schedule, with tasks placed automatically across three departments, zero scheduling conflicts, guaranteed by the solver itself. And you can see that same defect, right here, already placed on the Ambarnath–Badlapur route.
It's measurably smarter than random scheduling — the AI consistently prioritizes the most urgent work first.
Now watch this : I'll inject a brand new emergency defect, live. The AI scores it instantly and reoptimizes the entire schedule around it. And here's what matters most: every previously approved block :untouched. The system adapts to emergencies without ever breaking a commitment a human official already made.
This isn't a concept ; it's a working system, trained on real railway standards, tested and measured. BDMS solved how Indian Railways requests maintenance. Niyojan AI solves what it's still missing: deciding, intelligently, what matters most. This is Niyojan, by Team Rail Samanvay.
Â