What One Person Plus AI Can Actually Ship
I used to assume that enterprise-grade systems always needed a full team: product, design, frontend, backend, ops, QA, each person owning their own piece.
Building MakanCloud changed my mind.
MakanCloud is the first commercial product I've led from zero to launch. It's built for restaurants in Singapore and covers ordering, the central kitchen, delivery, staff scheduling, POS, and QR-code ordering. It's not a toy demo. It's a system that has to deal with real problems in real restaurants.
One person plus AI doesn't turn you into ten people
I'm not going to pretend AI can do everything. It can't understand what's actually happening on a client's floor for me, and it can't decide which feature should ship first.
But it has changed how I build.
When you build a system solo, the hardest part used to be context switching. One minute you're thinking through a business process, the next you're writing code. You finish an API, then you're chasing a bug. You fix the bug, and now you have to sort out deployment and docs.
Now I work with AI as a collaborator. It helps me break down tasks, fill in code, track down errors, and look back at how my workflow went. I still make the final calls and I own the results, but a lot of the steps that used to slow me down have gotten a lot shorter.
The real job is fixing business processes
Custom software isn't "let me build you a page."
This is especially true for restaurants and small businesses, where the problems tend to be very concrete:
- Order info is scattered all over the place
- The central kitchen and the outlets don't line up
- Deliveries only work because someone is watching them
- Staff schedules don't match what's actually happening in operations
- POS and ordering data never come together in one view
MakanCloud made one thing much clearer to me. Whether a system is worth anything doesn't depend on how long the feature list is. It depends on whether it cuts repetitive work, reduces back-and-forth, and brings the key data back into one place.
That's also what I care about most in custom development now: get clear on where the business is stuck first, then decide what system to build.
Enterprise-grade doesn't mean starting big
To me, enterprise-grade doesn't mean piling on complex architecture from day one, or building a heavy, bloated UI.
It comes down to a few simple things:
- The workflows have to actually run end to end
- When something breaks, someone has to handle it
- The system has to keep evolving
- The developer has to own the outcome
One person plus AI can ship a system like that, as long as the scope is clear and the priorities are realistic. You can't keep adding features just to look impressive.
I'll keep writing about how I use AI to build real products and custom software, including the mistakes, the trade-offs, and the failures. If you need a restaurant system, internal tools, automation scripts, or data dashboards, leave a comment or send me a DM and we can talk.