Loading…
.jpg&w=3840&q=75)
AI Engineering & OperationsBalkan Burger2026
From a Paper Ops Manual to a Restaurant Operations OS
I managed Balkan Burger in 2024 and wrote its operations manual by hand. Then, as Mirembe Muse, I rebuilt that manual into a digital-first restaurant operations OS — 60+ staff modules and three AI systems running across the group's branches.
+34%
table turnover increase
3 days
new-staff training (was 2 weeks)
60+
staff modules in the OS
3
AI management systems
The bottom line
Table turnover up 34% and new-staff training cut from two weeks to three days — the institutional knowledge captured once and deployed across every branch.
.jpg&w=3840&q=75)
The Challenge
Every manager at every Balkan Burger branch made slightly different decisions, trained staff in slightly different ways, and handled complaints with slightly different scripts. Across multiple branches the inconsistency compounded into a brand liability — and the entire operating system lived in people's heads and a paper manual. The work was twofold: first capture the institutional knowledge as a real manual, then turn that manual into software the floor actually uses, every shift, in real time.
The Manual, Written From the Floor
As the group's manager in 2024, I documented operations across all the branches — kitchen, front-of-house, and management — into a single operations manual, written in the register staff actually use, not the language of management consultancy.
It was the hardest kind of documentation: capturing excellent verbal knowledge without distorting it. Structured so any section stands alone — a new hire can read the guest-experience module without touching kitchen standards, and a manager handling a supplier dispute can go straight to the relevant section.
The Mothership: 60+ Modules, One System
Then, as Mirembe Muse, I rebuilt the manual into BB-MotherShip — a restaurant operations OS on a 74-table Postgres schema. Sixty-plus staff-facing modules cover shift boards and digital handovers, PIN clock-ins with automatic no-show detection, gamified training (XP, badges, leaderboards, streaks), and full back-of-house compliance logging — temperatures, oil changes, allergen briefings, waste, prep and production logs.
It even runs a customer-facing bread pre-order pipeline. And it was built multi-tenant from day one, so the same system scales from a handful of branches to a full franchise without a rewrite.
Three AI Systems That Manage While You Sleep
Three Claude-powered systems turn the floor's data into management intelligence. AI-scored pre-shift sales drills coach each role before service. The weekly "Black Book" is an AI management brief that synthesises attendance, incidents, morale, and shout-outs into one report. And an AI Competitor Radar tracks the rivals around each branch.
The result is a restaurant that runs on systems, not on any one manager's memory — the same operating advantage, captured once and deployed everywhere.
Watch it work
“Nanda is one of those rare gems who not only meets expectations but consistently redefines what excellence looks like — an unmatched ability to balance strategy with execution, with the emotional intelligence and natural leadership to rally a team around a shared vision.”
Bojan Ivanović
Co-Founder, Balkan Burger
Start a Project
Ready to build something like this?
Tell us about your project. We'll scope it, price it, and propose an approach — within 24 hours.