MOTHERSHIP
MAS Aqua

From phone calls and WhatsApp groups to a typed operating system.

MAS Aqua is one of Andhra Pradesh's leading shrimp seed hatcheries: a real technical business run by seasoned operators, over nothing but phone calls and memory. We embedded, extracted the tribal knowledge, and encoded it into a typed operating system with portals, atomic capacity, payment-linked orders, dispatch intelligence, and an agent layer on the same endpoints.

₹100cr+
hatchery business
4
user portals
86
API endpoints
18
system tables

How it works.

We do the unglamorous work first. Then the AI work becomes obvious, scoped, and shippable.

The business ran through humans.

A farmer does not place an order; he calls someone he knows by name. A distributor buys bulk capacity first and names his farmers later. Accounts confirms payments by checking the bank. Every step a human. Every handoff a place where the day can go wrong.

  • Orders living in phone calls and verbal distributor promises
  • Payments confirmed by manually checking the bank
  • Dispatch split across flights and trucks, coordinated from memory

We made the failure modes structurally impossible.

Capacity is booked atomically at the database level: two buyers, one remaining slot, the database decides and exactly one wins. Overselling stops being a discipline problem and becomes an impossibility. Orders confirm themselves when money moves, with advances and partial balances tracked cumulatively across transactions. Dispatch knows the difference between a flight and a truck, because they are different operations with different economics.

The production floor kept its Google Sheets. We made them a live output of the source of truth: workflow untouched, data finally trustworthy. And every action in the system carries who did it, because when livestock and lakhs move daily, the audit trail is how veterans trust a new system.

  • Capacity booked atomically at the database level
  • Prices and dispatch tied to shipping mode, advances tracked cumulatively
  • Google Sheets kept as a live output of the source of truth

Every actor got the interface they needed.

Farmers got Telugu self-service. Gujarat distributors got the two-phase bulk-order and farmer-allocation flow that exists in no software product on earth, because nobody who builds software has sat in that office. Admin got a single command center: one screen instead of five phone calls.

  • Telugu self-service for farmers
  • Two-phase bulk allocation for out-of-state distributors
  • Live admin view over capacity, payments, reallocations, and audit trails

The CEO can talk to the hatchery.

The dashboard is one client of the 86 endpoints. Agents are another. A company brain in WhatsApp answers the CEO from live state: what is blocking a delivery date, whose payment is pending, where cash stands.

  • Every dashboard action doubled as a typed API tool
  • Autonomous workers on the routine: collections, capacity protection, dispatch resolution
  • Telephone agents that answer farmers in Telugu and book the order

A non-digital company became machine-readable.

Orders take minutes instead of hours of phone tag. Double-selling is architecturally impossible. A CEO who asks his phone instead of asking around the office. The operators keep the judgment, the relationships, and the craft. The system carries everything else.

  • 219 production commits, 36 custom pages, live
  • 100% of actions attributed; the audit trail is how veterans trust a new system
  • The architecture ports: poultry, textiles, dairy, agri-inputs

Parallel proof of work

Explore how our core modules are deployed across different operational architectures.