HospitalityAI Services

Restaurant Floor Intelligence

CCTV computer vision that keeps live table state, balances covers and prevents double-seating.

Key Details

Challenge	A US startup building an AI restaurant-operations platform needed a live view of the floor from CCTV, tied to POS and reservations.
Solution	Computer vision on CCTV for real-time events, task routing, an operations assistant, POS and reservation data, and dashboards.
Technologies	Computer vision on CCTV, OpenAI API, AWS, Docker

Technologies used

Computer vision on CCTVOpenAI APIAWSDocker

Client

The client is a US-based technology startup building an AI platform for restaurant operations. It runs on cameras the restaurant already has, and combines that video with POS, reservation and staffing data. The reference describes it as an AI floor manager for full-service restaurants.

Product screenshots

Dark restaurant operations screen with an overhead CCTV view of a dining room where each table carries a coloured outline and state label such as ready to seat, needs bussing, check waiting or dining, a highlighted group waiting at the host stand, and a side panel with a waitlist, open alerts and per-section table counts. — Live floor view: existing CCTV is read to tag every table as dining, ready to seat, needing bussing or waiting on a check, and to spot guests waiting at the door.

Light cream floor-plan view of a restaurant with round and square tables coloured by state, staff avatars placed by section, a task queue routed to busser, host and server roles with timers and done buttons, and an operations assistant chat answering a manager's question about which section is overloaded. — Floor plan and task routing: table states from the cameras drive tasks to hosts, servers and bussers, and the operations assistant answers questions about the service.

What we built

Video analysis of existing CCTV to read floor activity
Event detection for tables ready to seat, guests waiting, tables ready to bus, and checks waiting to close
Task routing to hosts, servers and bussers
An operations assistant using staffing, sales and service data
CCTV insights joined with POS, reservation and staffing data
Live dashboards for floor state, alerts and management analytics

Engagement

Delivery is time and materials with 3 engineers and is ongoing.

Project team: 3 engineers. Duration: Ongoing. Model: Time & materials.

Outcome

LiveTable state updated

CoversBalanced

Double-seatingPrevented

Services used

AI Services →

Industries

Travel & Hospitality →