From notebooks to a phone in every hand

• Attendance went into a register: who came, who left early, who stayed late. Wages were worked out by hand from those pages.
• Labour payments happened whenever a worker asked, in cash or by UPI, in whatever amount was at hand. Who had been paid for which days depended on memory and loose pages.
• Harvest figures were noted on paper, if at all. Nobody could say with confidence how many kilos were in stock.
• Sales went out on credit. Each buyer’s running balance lived in the owner’s head or was pieced together from old messages.
• Expenses were scattered across receipts, bank messages and notes.

None of this is unusual. Most small farms and agri-businesses in India run this way, and they run well, as long as the owner is present, sharp and has time. The trouble starts when the business grows, when the owner is away, or when someone questions a number and nobody can prove it.

Labour was the biggest daily headache

Workers are paid by the day, at different rates for a full day, a half day and overtime. Payments don’t follow a monthly cycle: a worker might take ₹800 today against ₹1,000 owed, and the rest next week. Without a clear record it was easy to pay twice for the same days, or to forget a day entirely. A forgotten day means a worker goes unpaid, and on a farm that runs on goodwill, that matters.

Nobody trusted the stock number

Mushrooms are perishable. Some spoil every day, buyers return some, and the harvest changes from shift to shift. Without one reliable figure, the farm could promise a delivery it couldn’t fill.

Credit was hard to follow

Buyers took deliveries on credit and paid back in parts, by cash, cheque, UPI or bank transfer. Knowing exactly what each customer owed, and proving it with dated challans, took real effort.

Paperwork was done by hand

Quotations, delivery challans and invoices with the right GST treatment were drawn up one at a time, by hand or in a word processor.

Money in and money out were never in one place

Wages, purchases and customer payments were recorded in different places, when they were recorded at all. There was no single view of where the money went.

The owner was the bottleneck

Everything lived in one person’s notebooks and phone, so every question came back to the owner. Handing work to trusted staff meant handing over the whole notebook.

We read those messages closely, asked a lot of questions, and wrote every answer down as a plain business rule. Six of those rules shaped the whole system.

• Payments don’t follow the calendar. Labourers, staff and customers all work the same way: something adds to what is owed, a payment takes away from it, and the balance is what’s left. There is no month end, so the system runs on running balances, not monthly cycles.
• An empty day means absent, not forgotten. Sundays and holidays are paid only if someone worked. But a day nobody filled in shouldn’t quietly cost a worker their wages, so the app treats an unmarked day as absent and flags every date where attendance was never taken.
• A new rate never rewrites the past. When a labourer’s daily rate goes up, last month’s wages stay exactly as they were. Every day keeps the rate that applied on that day.
• Spoilage is normal, not an exception. Losing some produce every day is part of mushroom farming. The farm wanted production, wastage and returns as three separate numbers, so each one could be watched on its own.
• The work happens in the shed and beside the truck. Attendance is marked standing in a growing room; deliveries are recorded next to a vehicle. Everything had to work one-handed on a phone, in as few taps as possible.
• Trust is earned gradually. The owner wanted to hand specific jobs to specific people, one person on attendance and another on the harvest, without opening up everything else.

Labour: attendance and wages

• A daily screen to mark each worker as full day, half day or absent, with overtime added in 30-minute steps from a single button.
• Each worker has their own rates, and the app works out what they’ve earned.
• Pay any amount, any time, by any mode. If a worker is owed ₹1,000 and is paid ₹800, the app shows ₹200 still due.
• Pay for a date range: pick a from-date and a to-date, see every worker’s days, overtime, earnings and dues, and settle them in one go. Days already paid are skipped, so nobody is paid twice for the same day, and a paid stretch is locked for good.
• Dates with missing attendance are flagged, so nothing slips through.

Harvest and stock

• The day’s production in kilos, recorded by shift.
• Wastage recorded with a reason: spoiled, returned or other.
• One stock number, always on the home screen.
• A stock count: the owner enters what is physically there, and the app records the difference, who corrected it and why.
• A delivery can’t be recorded if the stock isn’t there, which keeps the number honest.

Customers, deliveries and payments

• Every buyer with one running balance, shown in red until it’s paid.
• Each delivery gets a date, kilos, a rate and an automatic challan number, and reduces stock at the same moment, so the two always agree.
• Payments in any mode, including part payments and advances.
• Quotations that go out before anything moves.
• PDF invoices, challans and quotations made and shared from the phone, with the farm’s details, the correct GST split for in-state and out-of-state buyers, the owner’s signature, a payment QR code and the amount in words.
• A full history for every customer: every delivery, every payment, every date.

Expenses and one ledger

• Quick expense entry against the farm’s own list of about 20 categories, which the owner can add to, rename and reorder.
• Wages appear as expenses automatically whenever a worker is paid, so nothing is entered twice.
• Older expenses can be back-filled, so the year’s records can be completed from the paper books.
• Every rupee in and out lands in a single ledger. Nobody types into it: it fills from the work people already do, and every entry links back to the delivery, payment or expense that created it.

Access and accountability

• The owner grants access person by person and section by section. A supervisor can start with attendance only, and get more as trust grows.
• Nothing is ever deleted. Workers who leave are marked inactive and keep their history; mistakes are corrected or cancelled with a reason; every change records who made it and when.
• Access can be switched off in an instant when someone leaves.

1. Secure access for the owner and their team.
2. Labour, the most painful daily problem.
3. Paying wages by date range, added after the owner tried the first version and showed us how they actually settle with workers.
4. Expenses.
5. Harvest and stock.
6. Customers, deliveries, invoices and payments.

The team was paying real workers through the labour section while later parts were still being built. That rhythm let the farm shape the product as they used it. Two examples:

• Paying over days already paid. The owner paid a worker for a couple of days, then tried to pay a wider week that included them, and got stuck. Now paying a range only pays the days not yet paid, and shows clearly which days it skips.
• Seeing spoilage on its own. The farm asked to see wastage and returns separately, instead of having spoilage quietly reduce the stock figure. We reworked the harvest section to track production, wastage and returns as three numbers.

Next on the roadmap is staff management for monthly-salaried employees. It plugs into the same running balances and the same ledger. Because the system was designed around how this farm actually works, each new piece slots in without changing anything the team already uses every day.

• Understand the business first. Learn the real rules, including the unwritten ones, before designing anything.
• Start with the biggest pain. Deliver the part that saves the most time and worry first.
• Ship early, listen, refine. Put real tools in real hands quickly, and let the people using them shape what comes next.
• Keep it simple for the person on the ground. Fewer taps beat more features.
• Make every number trustworthy. If the owner can’t trace a figure back to where it came from, it isn’t finished.