Restaurant-RGI

Your order guide is last spring's pars — your GMs override it or spend an hour a day rebuilding it, and one strange weekend still buries the walk-in or 86s your movers. Now multiply that by every unit, every day.

Builds and places the food & bev order itself — order / more / less / skip per category, tuned to this store's covers, weather and events — and takes over category by category only as it earns it.

A static order guide can't adapt and a recommend-only tool still needs a hand on every line. RGI is the one that places the order and hands back only the calls it hasn't earned — ordering you can actually leave on.

Auto-placed share climbs category by category; the gated line holds while the no-leash version overfits a noisy week.

You know what you bought — the invoices are right there. You don't know, by item by store, what reached the plate vs. what died in the walk-in or got over-portioned. The P&L shows the gap a month late, as one blurry point of food cost.

Lines up three numbers your systems keep apart — bought, should-have-used (sold × recipes), and actually-used (counts) — and names the item and the store where the gap lives.

Inventory suites record what happened; RGI closes the loop to what should have happened and acts on the gap — the half of food cost your invoices and your POS each only half-see.

Aggregate reads a healthy 30% while one item runs over; the same leak repeats across a cluster of stores.

A distributor quietly shrinks a pack, swaps a cut, or the case yield drops a few points. Your food cost creeps across a dozen stores and nobody can say why until month-end — if ever.

Tracks the realized yield of each item — usable product per case after trim and spoilage — across every store, and when it drifts below its learned norm names the item, the vendor and the week.

Nobody watches yield across stores — the leading indicator lives in your kitchens, not on an invoice. RGI catches the drift weeks before the P&L, and before the vendor admits it.

The salmon slips mid-stream; the flag fires the next week; the P&L wouldn't show it for a month — and it's the stores on one vendor lot.

Your 11-year distributor quotes +14%, "driver shortages." It clears as "the market" — because it's the vendor you trust most — and the overcharge disappears into a food-cost point nobody can take apart.

Scores every invoice line against your contract and the learned norm, decomposes the increase — commodity market move vs. supplier markup — checks it against your contract ceiling, and drafts the credit claim.

Others store what you paid; RGI knows why it moved and whether your contract allows it — and claims the difference at the line, not at the P&L.

The +14% splits into a market move within contract and a markup over the ceiling — the recoverable part.

You're in a buying group with cashback and rebates on lines you already buy — and a quarterly deadline you've never once hit, because matching SKUs and filing on time takes hours nobody has.

Matches your actual purchases to eligible SKUs and tiers, flags the in-group brand switches that capture cashback at no quality change, files before the window, and verifies the credit landed.

A rebate program is a live rule RGI holds and works, not a PDF in a drawer — and it learns which switches pay most, so capture compounds quarter over quarter.

Most of the eligible money is lost at "filed in time"; captured automatically, the rate climbs as it learns.

Eggs quadruple on bird flu, beef spikes, coffee blows up — or a heat wave plus a festival blows up a weekend. Most systems keep buying to par and you eat it, or you reprice a month too late.

When a volatile item's price (or the week's demand) breaks its normal range, it stops committing that item, switches to your backup spec, and flags which menu items to reprice or pull — then holds until it settles. On a demand shock it pauses the volatile categories back to your chef.

Every other system must always place an order. RGI is the only one that can say "trust me less right now" and cut its own exposure when the regime breaks — the capability nothing in the market has.

Buy-to-par gets whipsawed while Shield holds flat — because it knows when not to buy forward.

Every new store opens cold — the new GM over-orders for a quarter learning the room. And when a strong GM or chef leaves, the pars, vendor quirks and Friday instincts walk out the door. Across a chain, turnover and openings are a constant tax.

A new store opens already carrying what a comparable store learned — the pars, the vendor read, the daypart patterns — so the ramp is weeks not a quarter; and the judgment lives in the engine, so a departure doesn't reset the unit.

Rivals store order history and cost; RGI keeps the judgment and transfers it to a new store or a new GM. The continuity no competitor offers — turnover becomes a non-event.

A GM departure barely moves the line, and each new opening ramps faster than the last.

Prices get reset once a year by gut while costs move weekly; and the schedule is last week's copy-paste. Both halves of margin — the menu price and the labor line — ride the same demand your ordering already reads.

Price: propose +$ on the plowhorse where elasticity says volume holds, verify contribution margin on the next PMIX. Staff: size the shift to the covers forecast, cut the overstaffed rainy Tuesday.

The same learned demand model that orders your food prices it and staffs it — pricing and labor aren't two more tools to buy, they're one engine turned on the revenue and labor lines. roadmap

The plowhorse takes a price it can hold; labor tracks the covers curve instead of the copy-paste.

You order week to week and take whatever price and supply the market gives — a price-taker, exactly when prices swing most.

Because it knows how much of each ingredient you'll need and how sure it is, you commit the confident slice forward for locked price and guaranteed supply, and leave the uncertain rest to weekly buying.

A forecast you can't trust can't back a commitment. RGI commits only the slice it has earned confidence in — turning your demand into buying power without betting on guesses. roadmap

The committed share tracks confidence; the blended cost runs below taking spot every week.

The number you report up is food-cost %. Point tools give a one-time setup bump then fade — and can't prove the dollars to your CFO.

Every verified outcome sharpens the next, so food-cost % keeps trending down and recovered $ compound by source — waste, invoice, shrink, variance, rebate — each in a ledger you can hand finance.

A tool that runs Day-365 logic identical to Day-1 decays; RGI compounds — the gap against a static tool widens, never narrows — and every dollar is attributable, not a marketing estimate.

The cost-side line declines and holds while a generic tool drifts up; recovered $ stack by source.