Queue-pressure tablet for the floor lead
// What
One camera, one tablet, one number: 'how many people are in line right now and how fast is it growing?'
// Why
The floor lead already has authority; just give them a better window.
Change my answers
// Playbook · hospitality-bar
Hospitality with bar / queue-driven economics
Anywhere a queue costs you money, even modest measurement creates outsized returns. The trick is getting the measurement right and the response practical.
// Starter projects · sequenced
Sequenced. Sized. Honest.
From a confidence-building first step to a transformative bet.
MOMENTUM££3–6 weeks
Demand forecast + rota suggestion
// What
Combine your booking data + weather + local events into a 7-day rota suggestion. Staff lead overrides.
// Why
Saves the manager 2–3 hours/week and reduces over/understaffing.
TRANSFORMATIVE£££8–12 weeks
Closed-loop service signal
// What
Every covered/seated/served/paid event becomes data that retrains the staffing model and feeds the menu engineering team.
// Why
The combined signal is worth more than any single insight.
// Casework that grounds this
Built it for them.
CASE / 01
Sentinel — bar-queue & ambient retail watch
Ayr Pavilion (1,500-cap event venue)
Edge cameras + small object-detection model giving bar staff a live read on queue length, refusal risk and crowd density.
CASE / 02
Pavilion Festival — one CMS, 30 dynamic screens
Multi-stage festival operations
A single event-state JSON drives 30 cinematic signage layouts across 6 stages — change a set time once and it cascades everywhere.
// Risks worth flagging
- AI rota tools without override-discipline create resentment fast.
- Models trained on average weeks miss the weeks that matter most (Christmas, festivals).
// Watch out for in pitches
- Walk-away rate is harder to measure than people pretend; insist on a measurement plan before believing the ROI calc.
// The honest take
Most hospitality AI vendors over-promise on rota optimisation. The biggest reliable win is just 'see your floor in real time'.