A restaurant operations platform that forecasts demand, builds the schedule from it, runs the host stand, watches the kitchen, protects the margin and owns the guest relationship — one hosted system on one data model, with the numbers pushed to every manager each morning.
The platform
Tidal Exploration built DOME inside a working restaurant group, against real service, by people who had to live with the answer the next morning. Every module exists because something was breaking without it.
The platform
Most restaurant software treats forecasting as a report. In DOME it is the input everything else reads from: the schedule is generated against it, labor targets are enforced against it, purchasing envelopes are sized by it, and variance is measured back to it.
That means one number moves through the whole operation instead of four systems each holding a different version of next Tuesday.
Modules
Each module answers one operating question. They run on one data model, so the answer to each informs the next — kitchen pace moves table turns, table turns move covers, covers move the schedule.
A deep-learning model per venue, forecasting covers and revenue to the day and daypart. Not a moving average — it learns seasonality, holidays, calendar shifts and analog days.
A full scheduling system, not a labor report. Managers build against the forecast, targets are enforced before publishing, and the published schedule is what the clock enforces at the terminal.
The system a host runs the door on: reservations, waitlist, quoted waits, and a live floorplan showing which tables are seated, idle or turning — with the efficiency analysis coming free out of the same data.
Every ticket and cook event rolled up nightly, so kitchen conversations start from facts. Station load shows where the line stacks up and where plating gates the whole kitchen.
Leakage as a percentage of net sales, benchmarked against each venue's own prior-year baseline rather than a fleet average — and drillable to the check, the item and the approver.
Offers delivered by text to a guest who is already waiting or booked, then measured ring by ring against last year — with non-participating venues as a control group, so a lift is demonstrated rather than assumed.
A voice agent that answers the phone, gives hours and wait times, and books reservations directly into the floor system. Speech recognition, a language model, and layered voice fallbacks — all on platform infrastructure.
Mobile-first checklists completed on the floor and scored instantly: safety meetings, facility audits, and manager open, mid and close walks with photo capture.
Forecasting
Most forecasts hand a manager one number and leave them to decide how much to trust it. DOME produces quantiles — a low case, a median and a high case — so a schedule can be built for the likely night and checked against the busy one.
The model runs per venue rather than per group, because a beachfront room in July and a downtown room in February are not the same business with a different multiplier.
Through the month it corrects itself: completed days lock to actual point-of-sale mix instead of last year's ratios, so accuracy improves as the month unfolds rather than drifting.
Host platform
A host quoting forty minutes when the real wait is twenty loses the party at the podium. Quoting twenty when it is forty loses them at the table. DOME gives the host stand live table state, a quote built from what the floor is actually doing, and a text to the guest when the table is ready.
Because the host stand and the analytics are the same system, the measurement comes free. Every table's seated time, idle gap and quoted wait is recorded as it happens — so utilization is a by-product of running the door, not a separate reporting exercise.
Utilization is only credited under real pressure. A full floor at a slow hour proves nothing; a full floor with parties waiting is the number worth chasing.
In production
DOME is licensed to a multi-brand restaurant group operating across the United States. Every venue's numbers are pushed to its general manager and regional director each morning — reporting that arrives rather than reporting someone has to remember to open.
Guest engagement
A guest joins a licensee's loyalty program from a table-tent QR code or the venue's guest WiFi. They receive a membership card as an Apple Wallet and Google Wallet pass, carrying their member ID as a scannable barcode, the program's perks, and the venue's branding.
Passes are issued per licensee and updated when a program's terms change, so the card in a guest's phone reflects the current offer rather than the one printed the day they joined.
Consent is recorded against the exact text the guest was shown, stored in a ledger the application cannot alter, and honoured across every channel — because the record is what you rely on if anyone ever asks.
Consolidation
Multi-unit groups accumulate point solutions — one vendor for walks, one for phones, one for reporting, one for scheduling — each with its own bill, its own login, and its own view of the same restaurant. DOME is built to absorb those functions rather than sit alongside them.
Daily walks, safety meetings and facility audits run inside the platform, scored the same way across every venue.
An AI voice host answers, quotes waits, and books reservations — at a fraction of a per-venue phone contract.
Reporting is pushed from the same system that holds the forecast, the schedule and the cost, so there is nothing to reconcile.
The saving is real, but the reason to do it is that a single data model means an audit score can be read against the sales it affected, and a kitchen slowdown can be read against the turns it cost.
Questions
Delivery
DOME is delivered as a hosted platform. Each licensee gets its own deployment with its own data, configured venue by venue and reported across the portfolio — no shared tenancy of operating data between groups.
DOME reads from the point-of-sale and back-office systems a group already operates. The POS stays the system of record for transactions; DOME is where the operation is planned and measured.
General managers see their venue, regional directors see their region, and the executive view rolls up the portfolio — from the same system, with the same numbers.
The numbers are pushed every morning to the people who act on them, with monitors that raise an exception rather than waiting for someone to open a dashboard.
Tell us about your group and the systems you run today.