Menus, shifts and billing shaped to the concept, launch in days, and one system that follows the brand from pop-up to permanent — 16 experience solutions, narrated by Iris.
Brunch-only, 24-hour, pop-up, themed — the most creative formats in F&B usually run on the least suitable software, rebuilt by hand for every launch.
Ticket yield — what a seating earned against what it could have.
Run-sheet timing — whether each beat of the experience landed when it was meant to.
Pop-up payback — whether a site covered its setup before it was struck.
No-shows — the prepaid seat that never arrived, and what it cost.
Rather than a rebuild for every launch: one system that follows the brand from pop-up to permanent.
Menus, shifts and billing fit the format — a brunch-only window and a 24-hour room are not the same business wearing different hours.
A new concept, residency or pop-up is a configuration, not a project — which is what makes experimenting affordable.
One system across every site, so a concept that works in a pitch carries its history into a lease.
Ticket yield, run-sheet timing, pop-up payback and no-shows — live, with Iris naming the beat where the room lost the thread.
The same till in a permanent room, a festival pitch and a two-week residency — carrying tickets, merchandise and a menu that changes by the hour.
It trades without a connection. A field, a basement, a rooftop — service continues and reconciles when signal returns.
Tickets, food and merchandise on one check. A prepaid seating, a drink and a T-shirt settle as one transaction, not three systems.
Every sale knows its site and its night. Location and event tag attach at the line, so a pop-up's P&L exists from the first order.
Ticket yield, run-sheet timing, pop-up payback and no-shows on one live board — with Iris naming the beat where the room lost the thread.
Per concept, per site, per night. A residency, a permanent room and a festival pitch are compared on the same definitions.
The experience is measured, not just the revenue. Run-sheet adherence sits beside margin, because in this sector they are the same story.
Iris names the beat. Not that the night underperformed — which course, which act, which fifteen minutes it went wrong in.
P&L, balance sheet, cash and close on a single balanced ledger — with prepaid tickets recognised as they are consumed rather than counted when they sell.
Deferred revenue handled properly. A seating sold in March for a June night sits as a liability until the night happens.
Every site carries its own setup cost. Build, licence, pitch fee and strike are attributed, so payback is a real number.
Ask Iris for any number. Any concept, any night, any site — answered in the conversation, not a week after close.
Demand-planned replenishment for a business whose locations change — a fixed room, a festival weekend and a six-week residency each supplied against their own curve.
Planned per pitch, not per estate. A festival weekend has no history at that site, so the forecast borrows from comparable events instead.
Ticket sales are a demand signal. Prepaid covers are known volume, so a themed night is stocked to a number rather than a hope.
Load-out is part of the plan. What travels, what returns and what is written off is planned rather than discovered in a van.
Digital checklists, shift routines and standard operating procedures — and in this sector, the run sheet itself: the sequence of beats the experience is made of.
Build and strike, signed off. Setting up a pitch and breaking it down are procedures with evidence, not improvisation with a van.
Every beat is timed. Run-sheet adherence is recorded per seating, so a night that drifted has a timestamp rather than an opinion.
Exceptions escalate themselves. The routine that didn't complete surfaces to a human; the rest stays quiet.
Rostering, attendance and labour percentage measured against forecast demand — for concepts whose shifts do not fit a standard working day.
Overnight shifts handled properly. A shift that crosses midnight is one shift, correctly costed, not two half-days and a reconciliation.
Event crews rostered separately. A festival weekend draws on a different pool with different rates and travel, planned as its own thing.
Night premiums built in. Unsocial-hours rates are part of the labour model rather than a payroll surprise.
Deterministic logic under the hood (the “Zen Rules” layer) — and in a sector that trades in fields, basements and rooftops, the most important rule is that trading never stops. Three solutions run on this layer.
Trading has continued uninterrupted — 213 transactions queued locally. No guest has seen an error. Sync will reconcile on reconnection with conflict rules already applied.
What the staff member experiences when there is no signal: nothing. The till behaves identically, the guest sees no difference, and no one has to decide what to do about it.
Orders, payments and tickets are taken exactly as they would be online — there is no separate offline workflow to remember.
Card payments are captured for deferred authorisation within configured risk limits, so a lost connection is not a lost sale.
Menus, prices, ticket lists and guest records are cached locally, so an entire service can run from the device.
Ticket validation works offline against the cached seating list, which is exactly when a festival gate needs it most.
The connection state is visible to a manager but never to a guest, and the till never asks staff to make a judgement call.
The architecture underneath — how several devices that were all offline at once converge into one correct ledger, without duplicates, lost orders or a manual reconciliation.
The system is offline-first by design rather than online-with-a-fallback, which is the difference between graceful and fragile.
Every transaction carries an idempotent identity, so a retry on reconnection can never post the same sale twice.
Conflicts — two devices selling the last ticket, stock depleted twice — resolve by defined precedence rather than by whichever synced first.
Sync is incremental and resumable, so a device on a weak signal converges rather than repeatedly failing a full upload.
Anything that could not resolve automatically is queued for review with its context attached, rather than silently dropped.
Ticketing platforms, payment providers, local tax and the accounting ledger — connected and watched, and configured per market so a concept can travel without a rebuild.
Ticketing and booking platforms connect through one hub, so a seating sold on a third-party site is the same seating in your system.
Payment providers vary by market and by pitch, and the hub abstracts that so the till does not care which one is in use.
Tax rules are configured per market, which is why taking a concept to another country is a setup rather than a release.
Sales, ticket revenue, refunds and settlement post themselves to the accounting system, already coded to site and event.
Every connection is health-checked continuously, and a degraded ticketing API is flagged with the seatings at risk.
Forecasting, optimisation, attribution and scoring models under the hood (the “Zen Models” layer) — because a concept with no two nights alike cannot be run from last week's numbers. Twelve solutions run on this layer.
When a seat is sold in advance, the restaurant becomes a venue. Seatings are released, priced, filled and protected against the no-show that a prepaid model makes expensive.
Seatings are released as inventory with capacity, timing and price tier, rather than as a reservation book with money attached.
Yield is managed across the release — early tiers, last-minute pricing and hold-backs are set against forecast fill.
No-show risk is predicted per seating, so overbooking buffers are set deliberately rather than avoided out of fear.
A confirmed seating enters the run sheet and the prep plan directly, since prepaid covers are known volume.
Waitlist releases fill cancellations automatically, which is where most of a sold-out night's lost revenue actually hides.
The money side of a ticket — taken months early, recognised on the night, refunded under policy, and split with whoever you collaborated with.
Ticket revenue is held as deferred income and recognised on the night it is consumed, not on the day it was sold.
Refund and transfer policy is applied by rule — cut-offs, partial refunds, credit rather than cash — without a negotiation per guest.
Collaboration and revenue-share splits are computed automatically, so a guest chef or a venue partner is settled from the same ledger.
Deposits, balances and on-the-night spend resolve into one guest position rather than three separate records.
Cash timing is modelled across the release window, which matters when a concept is funded by its own presales.
The night as a sequence of beats — course, act, reveal, reset — scheduled, staffed and timed. In an experience concept the run sheet is the product, not the paperwork.
The run sheet is built as a timed sequence with kitchen, bar, front of house and performance cues on one timeline.
Optimisation resolves the constraints — oven capacity, staff movement, reset time — into a schedule that can actually be executed.
Beats fire to the right station at the right moment, so nobody is watching a clock and guessing.
Actual timing is captured against planned, giving a per-beat record of where a night ran long or short.
The sheet is versioned and reusable, so a proven night can be run again or handed to another site exactly.
Brunch-only, late-night-only, Sunday-only. A menu that exists for three hours is a discipline problem as much as a scheduling one — so the lockout is enforced rather than remembered.
Menus open and close on schedule across till, app, web and every aggregator in the same second.
Items outside their window cannot be rung through, so the concept holds even on a manager's day off.
Clustering identifies which items genuinely belong to a daypart and which are being sold out of habit across all of them.
Windows are forecast against demand, so a brunch that consistently sells out at 11am is a case for extending it.
Prep and par levels follow the window, so a three-hour menu is not stocked as though it ran all day.
Most hospitality software assumes a day ends. A 24-hour room has no natural close — so the business day, the cash-up and the reporting boundary all have to be defined rather than assumed.
The business day is configured to the concept, so a 4am trading hour lands in the night it belongs to rather than the morning after.
Cash-up and reconciliation happen on a rolling basis without stopping service, because a 24-hour room cannot pause to count.
Overnight labour is costed correctly across the midnight boundary, including premiums, as one continuous shift.
Demand is forecast across the full 24 hours, so the quiet hours are staffed to their real curve rather than to a minimum.
Late-night margin is reported in its own right, since it behaves nothing like the daytime business it shares a room with.
Planning and running the programme — which nights to hold, who to collaborate with, how long a limited run should last, and how to launch one without rebuilding the system each time.
A themed night is configured as a package — menu, pricing, capacity, staffing, run sheet — and cloned for the next one.
Demand is forecast before release, so capacity and pricing are set against expected fill rather than optimism.
Segmentation identifies which part of the audience each concept actually attracts, which is what makes a programme rather than a series.
Collaboration terms are held with the event, so a guest chef's split and billing are configured once and executed automatically.
A limited run has a planned end, so a concept is retired deliberately rather than quietly outstaying its audience.
What a night actually earned — separating genuinely incremental trade from the regulars who would have come anyway, which is the only honest measure of a themed programme.
Attribution separates incremental revenue from displaced revenue, so a busy Tuesday that emptied Wednesday is seen for what it is.
Full cost of the night is carried — collaboration fee, extra labour, marketing, bespoke stock — not just food cost.
New guests acquired are tracked beyond the night itself, since a themed event's real return is who comes back.
Nights are compared on contribution per seating rather than on how full the room looked.
Iris recommends which formats to repeat, which to rework and which to stop, with the modelled effect of each.
Where to trade, and making sure every transaction knows where it happened. Without the tag there is no per-site truth — and without per-site truth a pop-up programme is guesswork.
Every sale carries its location and event tag automatically, so a site's performance exists without anyone reconciling it later.
Sites and pitches are scored before committing — footfall, audience fit, pitch cost, comparable past events.
A new pitch inherits a forecast from comparable sites rather than trading blind on its first weekend.
Recurring venues build history, so returning to a festival is planned from what it did last year.
Sites are ranked on what they actually returned, which is what turns an opportunistic calendar into a route.
A full P&L per site and per run, including the costs a permanent restaurant never has — build, licence, pitch fee, transport and strike. Payback is either real or it is a story.
Setup, licence, pitch fee, transport, crew travel and strike are attributed to the run that incurred them.
Payback is tracked within the run, so you know on day three whether a two-week residency will cover itself.
Sites are compared on contribution per trading day, which is the only fair comparison between a weekend and a season.
Shared central costs are allocated on a defensible driver, so a pop-up is neither flattered nor unfairly burdened.
A pop-up's history carries into a permanent site, so a lease decision is argued from its own trading record.
Running several pitches at one festival, or several festivals at once — a distributed operation with no back office, no reliable signal and a weekend to get it right.
Every pitch reports into one live view, so an operator sees the whole weekend rather than phoning each van.
Stock is forecast and allocated per pitch, with transfers between pitches recorded rather than improvised.
Demand is forecast against event programming, because a headline set empties a queue and then floods it.
Crew, kit and vehicles are scheduled across overlapping events, so two pitches cannot be promised the same fryer.
The whole operation runs offline-first, which at a festival is the normal condition rather than the exception.
In this sector the audience is the asset. A waitlist, a membership and a community list are what let the next concept open to a full room on day one instead of an empty one.
The waitlist is a demand signal, not a queue — its size and speed tell you what to price and how much to release.
Membership and early access are managed as products with their own economics rather than as a mailing list.
Segmentation identifies which parts of the community follow which concepts, so a launch is announced to the right people.
The audience follows the brand across sites and concepts, which is what makes a pop-up programme compound rather than restart.
Lapsing members are identified while the habit is recoverable, since a scarce-access concept loses people quietly.
Which concepts deserve to continue — scored on contribution, audience response and sentiment together, because a beloved concept that loses money and a profitable one nobody talks about are different problems.
Every concept is scored on contribution, repeat rate, fill rate and sentiment as one composite rather than four disconnected reports.
Sentiment from reviews and social is themed automatically and read against the commercial numbers, not beside them.
Clustering groups concepts by what actually drives their performance, which is how a repeatable formula is found.
Early signals are surfaced within a run, so a concept can be adjusted mid-flight instead of judged at the end.
Iris recommends scale, repeat, rework or retire — with the modelled effect on the wider programme.
Not a screenshot — the actual agent. Natural-language answers across every concept, site and night. One solution runs on this layer.
The margin an experience concept leaves on the table — the record, the hot sauce, the chef's table upgrade, the next residency's ticket. Offered at the moment the guest is most inclined to say yes.
Merchandise and retail lines sit in the same catalogue and the same check as food, rather than in a separate shop nobody reconciles.
Bookable extras — upgrades, pairings, kitchen tours, private tables — are offered at the point in the journey where they actually convert.
Iris times the offer to the beat of the night, because a memento sells at the end of an experience, not at the booking.
The next concept's tickets are recommended to the guests whose history says they will want it, turning one night into a programme.
Retail stock, packaging and fulfilment cost carry into the line, so merchandise margin is real rather than assumed.
Sixteen solutions. Every card opens a live guided demo.
You configure a sector playbook, not a custom project. In a sector that launches a new concept every few months, that distinction is the whole proposition.
Onboarding is agentic. Iris connects the till, the ticketing platform and every payment provider directly — no manual data mapping.
A new concept is a clone, not a build. Menu, run sheet, pricing and staffing are cloned from a proven format and adjusted — which is what cuts launch time.
Pop-up to permanent, same system. A concept carries its trading history into a lease rather than starting from zero.
The first line of support is agentic — Iris resolves most of it herself. Our engineers pick up from there.
Layer 1 — Iris, 24/7. Which matters more here than anywhere: a concept trading at 3am cannot wait for office hours.
Layer 2 — our engineers. Anything Iris can't close escalates automatically to a Zentallio engineer.
No blank tickets. Every escalation arrives with Iris's own diagnosis — engineers start from an answer.
“Brunch-only, 24-hour, pop-up, themed — the most creative formats in F&B usually run on the least suitable software, rebuilt by hand for every launch.”
Zentallio ships the playbook: menus, shifts and billing shaped to the concept, launch in days, and one system that follows the brand from pop-up to permanent. Underneath sits the same platform every sector runs — one data spine from the till to the ledger, three layers of intelligence, and an agent that narrates every screen, flags what needs a human, and acts on the rest.
A themed night that filled the room by emptying the following Wednesday, counted as a success.
A pop-up whose setup, transport and strike costs were never attributed, so payback was a feeling.
Every launch rebuilt by hand, so the cost of trying something is the reason you stop trying.
Trading never stops — offline-first by design, with sync that converges without duplicates.
Seatings, run sheets, site P&L and concept scoring are modelled — twelve of the sixteen solutions.
“Where did the room lose the thread?” — answered against the run sheet the night was built from.
Every Zentallio sector runs the same three-layer intelligence, tuned to how that format actually loses margin. This is the last of the ten.