LMRD Hospitality: AI-assisted inventory management web app

Case study: an AI-assisted inventory platform for LMRD Hospitality with multi-branch stock, purchasing and wastage tracking, RBAC and Postgres RLS.

About the client

LMRD Hospitality: Hospitality.

The project

An AI-assisted, full-stack, responsive inventory management platform built for LMRD Hospitality to run stock, purchasing, and wastage tracking across multiple branches from one system — with authentication and role-based access control (RBAC) for Admins and branch Managers, backed by Postgres Row Level Security as the real authorization boundary behind every permission check.

Type: Web App + AI Solution. Built in 14 days, live at inventory.lmrdhospitalitycontrol.com. Tech: Next.js, Supabase, PostgreSQL, React, TypeScript, Tailwind CSS.

Key features

• Role-Based Branch Access: Admins manage every branch while Managers are scoped to exactly one — enforced by Postgres Row Level Security, not just app-level checks
• Purchase Request Workflow: Managers raise purchase requests with vendor and quantity details; Admins review, approve, revise, or reject across all branches from a single queue
• Real-Time Stock & Reporting: Consumption and wastage logging, low-stock alerts, cash-flow insights, and downloadable PDF reports — all generated live from current stock data

Results

LMRD Hospitality moved off manual, spreadsheet-based stock tracking and onto a single real-time system across all its branches. Managers now get instant visibility into stock levels and low-stock alerts without manually checking thresholds, purchases flow through a clear request-and-approval chain with invoices attached and stored automatically, and ownership gets a live, branch-by-branch view of purchase spend and cash outflow for the first time. The entire platform was scoped, built, and shipped in 14 days.

Before this, every branch was tracking stock on its own spreadsheet, and getting a straight answer on what we actually had on hand meant calling around. Now it's just one system, and I can see every branch's stock, every purchase, every bit of wastage without chasing anyone down. The approval flow on purchases has also made it a lot harder for spend to slip through without someone signing off. For something that touches this much of how we run the business, fourteen days to build it was quick.