Restaurant Order & Billing App
Product Design, AI, Shipped
2024
Founder & Lead Designer
Background
Mewari Makkhan Dhaba is a family-run restaurant on the highway near Udaipur, with an AC room, a Non-AC room and seating for up to 80 guests. I had already built their QR menu, so when Priyank, the owner and my childhood friend, needed a billing tool, he called me.
For 12 years, my work ended at the design handoff. This project was the first time I designed, built and shipped a product end to end, using AI as my coding partner..
THE PROBLEM
His ask was simple: "Koi app ya software batana, bill banane ke liye, simple sa." As we talked, kitchen slips (KOT) came up too. Every order lived in a diary at the counter, and whenever staff needed to confirm an order, he had to walk back and check it.
Ready-made restaurant apps charged a monthly subscription and came with dashboards he didn't need. So the brief was clear: no subscription, works offline, one phone and a small 58 mm printer, and no training for a small, busy team.
THE PROCESS
Before any screen, I mapped the life of a single order, from the moment a guest sits down to the moment the table is free again. Three people touch it: the waiter, the kitchen and the owner. Every screen in the app had to move an order one step forward.
AI's First Version, and What I Removed
I described the idea to Claude in a few lines. No perfect prompt; I just wanted to see if I could get it running on a phone. Fifteen minutes later, it was, and it came with things nobody asked for: staff logins, PINs, owner and waiter roles, tax settings, six tabs and eight settings menus.
It looked impressive, but it was far too much for a family team. AI never asks "does he really need this?" That question is still the designer's job. I spent the next 30–40 minutes cutting: logins and roles went, six tabs became one home screen, and eight settings menus became two.
Designing the Flow
Taking an order follows the map in three steps: pick the table, tap the dishes from the real 105-dish menu, then check quantities and add kitchen notes. The home screen shows every running order with its status, bill, waiting time and print state, and dishes are ticked as they reach the table.
Printing was tested on the restaurant's 58 mm thermal printer. The kitchen slip shows the table and dishes in large type with no prices, and reprints are clearly marked so nothing is cooked twice.
Shipping Feedback the Same Day
I shared the app with Priyank, and he used it that same evening. The next morning he said: "Perfect, aisa hi kuch chahiye tha." When I asked how many orders he had recorded, he asked back: "Kya ye bata sakta hai kal kitni sale hui? Kitna cash, kitna UPI?"
Every paid order was already saved with its payment type, so it only needed one screen. I added it in 10–15 minutes, on the same call. That evening he updated the app and called back, excited: "Yaar, wo feature bhi aa gaya!"
The Impact
The restaurant now runs on the app every day. Orders reach the kitchen on a printed slip, every table is visible at a glance, and at closing Priyank knows the day's sale, cash and UPI. No diary, and no monthly fee.
For me, it changed how I see my role. AI made building easy. Choosing what really matters is still on us.
Next Step
The app is mobile-first by design, so the restaurant didn't need to buy a counter terminal or monitor. If the restaurant grows, the next steps are a tablet layout for the counter, syncing orders across devices, and a kitchen screen view, each added only when the restaurant needs it.
