Ask anyone who lives in a housing society what goes wrong. It's rarely the lift. It's the water pump that dies at 6 a.m. or the generator that won't start during a power cut.
That's the opening for a property maintenance business that warns before things break. I'll work through it using one idea, Humcheck.
The problem
Most societies fix pumps and generators after they fail. Nobody watches how they run day to day, so the early signs, like a pump drawing more current than usual, go unnoticed. When it fails, residents get dry taps or dark nights and the committee gets angry phone calls.
How the business works
Humcheck puts clamp-on current sensors and a small vibration sensor on a society's water pumps and diesel generator. It learns how each machine normally runs and warns the facility manager when that changes.
The clamp-on part matters. Nobody has to rewire anything, so a technician can fit it in an hour and the committee doesn't have to approve a big job.
Who pays
Facility-management firms pay per machine each month. That's the smart buyer, because one firm runs many societies and wants fewer emergency callouts. The society benefits, but you don't have to sell to every committee one by one.
What it costs
Our estimate for Humcheck is ₹60 L–1.3 Cr over 6–9 months. The money goes on sensors, gateways, the model that learns normal behaviour and field technicians.
Four nearby ideas give you a sense of range:
- Driptight, data leak control for small CA firms, at ₹45–85 L over 5–8 months.
- Commonhour, hosted coworking days for remote workers in smaller cities, at ₹30–60 L over 3–5 months.
- Mitelow, a measured dust mite clean-up for allergy patients' bedrooms, at ₹45–80 L over 4–6 months.
- Yonderlet, rental management for owners living abroad, at ₹45–90 L over 4–6 months.
Yonderlet is the closest cousin. It's property maintenance sold to the owner, not the building. If you like Humcheck then Yonderlet's lesson is worth taking - one missed task, like a quarterly tax return, can end the referral chain.
Licences
No sector licence is needed to run a monitoring service, but the hardware has rules.
- WPC equipment type approval from the Department of Telecommunications for each gateway model with a cellular, Wi-Fi or Bluetooth radio, before you sell it.
- Electrical safety for fitting. Clamp-on sensors don't touch the wiring, but have a licensed electrician sign off installs where your state needs it [Requires validation].
- The baseline: company or LLP registration with the MCA, and GST once you cross the threshold.
- DPDP Act if you store any resident contact data for alerts.
The first 90 days
Days 1–30. Talk to five facility-management firms. Ask how many pump and generator failures they had last year and what each callout cost. If they can't tell you then that's a sign the problem isn't priced yet, which makes selling harder.
Days 31–60. Fit sensors in three societies run by the one firm that was most interested. Don't send any alerts yet. Just learn what normal looks like.
Days 61–90. Turn alerts on. Count the warnings that came true, the false alarms and the failures you missed. That's your proof. If you caught even one failure early then the firm has a story to tell its other societies.
How AI fits
AI learns each machine's normal pattern. That's the core of the product. But don't trust it to say "this pump will fail on Tuesday". Say "this pump is running differently, please check". A soft, honest alert gets acted on. A confident wrong one gets the system switched off.
The risks
- False or missed alerts. One loud false alarm or one silent failure can end a pilot.
- Budget resistance. Committees cut spending that shows no visible result in a few months.
- Copycats. Society management apps could add basic monitoring.
- Field quality. A sensor fitted badly gives bad data. Train your technicians properly, check every install with a photo and a test reading, and keep a spare kit in each city so a failed sensor is replaced within a day. Cheap hardware with slow service is how monitoring businesses lose trust.
My take
This is a good property maintenance business because the failure is visible to every resident and the fix is cheap to fit. Sell to facility-management firms, not committees, and price it below the cost of one emergency callout a year. If you do that then the sale almost makes itself.
For more, see the real estate ideas and the India idea hub.
Our estimate for Humcheck is ₹60 L–1.3 Cr over 6–9 months, mostly hardware, the learning model and field staff.
Do I need a licence to monitor society pumps?
No sector licence, but each wireless gateway model needs WPC equipment type approval.
Who should I sell to?
Facility-management firms that run many societies, because they pay for every emergency callout.
Talk to Yash
If you're planning a property maintenance business and want to design a pilot that proves value in 90 days, I can help you plan it. Talk to Yash →
For illustration only. Ideas are hypothetical and names fictional; sourced facts may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.