The dispatch dashboard
Spanish-first ยท Multi-language
Orders arrive from Mercato-hosted ecommerce and CompuDime POS. There is no unified view. Both systems require separate logins and training.
The communication between online order picking and delivery departments is conducted via radio, WhatsApp, and phone calls. This leads to inefficient and siloed messaging across the store.
Delivery context such as gate codes, bag types, and timing is inputted in a free-text note field typed by a cashier at the completion of purchases. This can cause the notes not to stand out on the delivery label, causing employees to miss important information at crucial points in the delivery process.
Routing decisions live entirely in the delivery manager's head.
Missing bag at scan time: drivers know a box is missing, not what it is or where it is.
Delivery, FedEx orders, pickups, and Grab & Go do not share a common system. Everything is managed manually.
In-store POS transactions and online Mercato orders in one queue. The pipeline strip shows incoming orders before delivery labels print.
The system is neighborhood-aware, weather-triggered, and packing-time-aware. The system suggests that the manager send a specific route sequence to a driver and the manager confirms.
BP1, BP2, FT, MAN, KEN, every order is tagged from address. For example, the algorithm treats deep Borough Park differently from surface Borough Park.
The app uses turn-by-turn voice navigation, directs drivers to use a LIFO loading sequence, and interacts with drivers using gamified elements such as points for each scanned box and bonus points for routes finished quickly. Shelah is an app built for the actual workforce.
Vans cannot depart until every box and frozen bag is confirmed by scanning each box.
Separate views for every order type. FedEx countdown clock. Pickup date filtering. Each category is managed without confusion.
Next.js + Supabase + Vercel. Supabase uses WebSockets to provide instant real-time sync between the dashboard and the driver app.
Converts cashier delivery remarks (unstructured data) into structured data the system can use, enabling intelligent delivery routing before the print module is built.
Virtual printer driver intercepts CompuDime print jobs. Generates Shelah-formatted labels. Shelah owns the label, enabling refined staging intelligence.
OpenWeatherMap API. Alerts trigger above manager-set threshold (default 70ยฐF). Cold van assigned to Manhattan first. Borough Park run โ add ice packs to fridge boxes.
Orders entering the system from Mercato or CompuDime appear in the pipeline strip above the dispatch queue before a label is printed. Once the label prints, the order moves automatically into the dispatch queue. The manager sees what is coming before it arrives.
Spanish + English in the driver app and dashboard from day one. Yiddish, Hebrew, and Russian added to the manager dashboard in V1.5. Spanish and Yiddish are competitive differentiators, because Scanovator does not offer either.
The app is a tool built for the driver's language and workflow. The inclusion of gamification elements instills a sense of pride and accomplishment in the drivers. The app also makes it easy for drivers to track their customers' tips and to transfer them to their bank account.
Deliveries will be on time with instructions reliably followed because the algorithm will use customer-chosen timeframes as a variable in suggesting routes to be sent to drivers. The alert banner that is triggered by geofencing ensures drivers see important notes before leaving the van, creating more satisfied customers and reducing their complaints.
Hybrid in-store and delivery operations using CompuDime POS and Mercato. Pain points: labor cost, delivery failures, manager burnout. Decision maker: owner or store manager.
Ethnic grocery stores servicing customers via 50+ deliveries per day with the same workforce demographics and tribal knowledge problem.
Specialty retailers nationally that share the same in-store and online delivery fragmentation.
We have already purchased the domain name Shelah.ai through Cloudflare. Following that we need to setup accounts with Supabase, Vercel, Cursor, Google Play Store, Apple App Store, and Google Workspace.
Build the system using vibe coding technologies such as Cursor. Pilot the system dashboard and driver app at Pomegranate Supermarket and in the roads of Brooklyn. Set up QuickBooks, a business account, Stripe for payment processing, and file as a C-Corp in New York State.
Onboard Kosher supermarkets in the Tri-State area.
Hire a salesperson to focus on support and onboarding additional stores including ethnic groceries in the Tri-State area.
Hire a developer to support the system and fix bugs as they occur. Target specialty and ethnic retailers located in densely populated cities across the United States.