Lucky Shovel builds AI systems for real-world operations — systems that answer, remember, watch, and decide. We measure the baseline before we build, and we show you exactly what changed.
It's trapped in WhatsApp threads nobody can search. In the head of one overloaded owner. In documents scattered across clients and chats. In repetitive questions, forgotten follow-ups, and field signals nobody reads in time. We build the system around that trapped capability — so it can move again. The fuller argument — why I build this way — lives at zzag.net.
Customers ask. Nobody replies in time. The lead cools, the booking slips, the owner apologises again.
Client history, promises, prices, decisions — all carried by whoever has been there longest. Usually the owner.
The contract is in email, the amendment in WhatsApp, the proof of address in someone's camera roll.
The data exists — in the field, in the logs, in the inbox. It just never becomes a decision in time.
We think AI should be measured by what it changes, not what it promises. Every engagement runs the same loop — and the loop has an exit.
We sit inside the real workflow first — people, tools, bottlenecks, failure modes. No deck. No demo theatre.
Response times. Missed leads. Manual hours. Open obligations. If we can't measure it, we won't claim it.
Inside the tools your team already uses — WhatsApp, Telegram, the phone line. Not another dashboard to ignore.
Usage, behaviour, outcomes, drift. Every meaningful action the system takes leaves a receipt.
If the numbers move, we scale it. If they don't, we say so and stop. You should never pay to keep our experiment alive.
Problems that kept repeating across clients became focused products. Each one stands on its own; all of them run on the same method.
Supervised WhatsApp text replies, owner-review handoffs, and tentative Google Calendar holds. For Singapore service businesses testing one valuable job.
● Design-partner pilot · S$300 / four weeksPrivate operating memory for owner-led teams. Chat mess in — client context, open loops, receipts, and finished work out.
● Pilot deployments openGets Singapore SMEs audit-ready across PDPA, safety, and employment — starting with a free assessment. Runs on MRDK.
● In active client workPlantation intelligence for oil palm estates — digital twins and predictive analytics, built with operators in the field.
◌ In field developmentOur systems issue a receipt for every action they take. Our engagements issue a receipt for every outcome they claim — same rule, all the way down. Completed receipts stay separate from evidence boundaries; where an outcome has not been measured, we say so.
"Every business owner I know needs to see this."— Shawn, Infiniti Professional Services
Disclosure: Zero Wait is founder-run and dogfooded on its public enquiry line. No Fit Kaki or buyer outcome receipt is claimed for Zero Wait.
Systems builder. Also co-founder of Fit Kaki, a senior-strength social enterprise — because trapped capability isn't only a business problem. zzag.net
Makes sure that what was promised is what gets delivered — and that every engagement leaves a paper trail worth auditing.
Tell us where the work hurts. We'll tell you honestly whether a system would help — and what we'd measure first.