net sales, daily, all three locations · day-of-week x heat x events
ActualForecast80% bandToday (projected)
Forecast runs on the September run-rate, not July. Stone crab opens Oct 15, and the forecast only sees it if someone feeds it the season.
Today · Sun Sep 27
projected close · snapshot date
district net sales
Camelback / rest of district
Projected guests
Avg check · brunch vs dinner
Peak windows
vs same day last week
Foot Traffic · Today by Hour
checks opened vs outdoor temp
Checks openedOutdoor temp (right axis)
Happy hour sits inside while the patio stays empty until the heat breaks after 6. Staff to that, not to the clock.
Location Comparison · Last 7 Days
the view you open on Monday morning
Location
Net sales
WoW
Guests
Avg check
Labor %
Sales / labor h
Perishables · Projected Outflow
on hand vs 7-day usage ratecover = days of stock at current usage
Item
Category
On hand
7-day outflow
Cover
Status
Toast has no ingredient or purchase feed. The case-level view is built from your menu, portion weights and close-out counts. That build is the product.
Staffing vs Sales · Next 7 Days
district schedule vs model recommendation
Role
Scheduled h
Recommended h
Delta cost / wk
Rate
District scheduling flags from the model
Waste and Shrinkage · This Week
expected vs actual outflow
Expected (sales x recipe)Actual (usage + close-out)
Profit Forecast · Next 4 Weeks · District
stacked: cost of goods, labor, fixed · green = projected gross profit
Assumes current cost card, 25.2% labor, $38,500/wk fixed. Forecast landed within of actual over 28 days. Not GAAP.
The Order Message
arrives 6:00 AM the morning you place the order
Quantities are computed here: usage through the next delivery plus one day of safety, minus on hand. Nothing sends in this demo. In production, Accept writes one row to the decision log, and that row is what makes the savings provable twelve weeks from now.
Monday 7:00 AM · The District Memo
forwards to ownership unchanged
Same numbers as every panel above, written out in the order you would say them. This is the part that gets you renewed, because it is the part your boss reads.
Where Each Number Comes From
what Toast gives us, and what we have to build
Panel
What we pull
From where
District Sales and Forecast
84-day daily net sales, guests, order counts, grouped by location and revenue center
/era/v1/metrics + one-time /ordersBulk backfill
Foot Traffic
Per-check open time by hour, dining option (patio vs main), server
/era/v1/check
Staffing
Scheduled shifts, time entries, hours and cost by job, sales per labor hour
/era/v1/labor + /labor API (shifts, employees)
Perishables
Stock status and counts as they change, plus usage built from your menu, portion weights and close-out counts
/stock API + webhook + the layer we build
Waste and Shrinkage
Item waste counts, excess-food orders, expected vs actual ingredient outflow
Quantity sold per item against your cost card, labor, fixed cost by site
/era/v1/menu + your cost card
This page runs on simulated data.
Every number above is generated. What is real is the shape of it. Point it at your Toast and the same panels run on your numbers. When Naples or Houston opens, the district grows by one restaurant ID, no rebuild.
Scripted demo. Every answer is computed from the same simulated data as the panels above. The production build puts an LLM interface on a vetted metric layer: it never invents a number, and it refuses what it has not been given.