Capacity is already committed.

By the time the morning proves itself, capacity is usually already spoken for.

The roster went out. The shift is staffed or short. The dock window is booked. Machine hours are allocated. Goods spend is ordered against a hypothesis that has not finished proving itself. Demand is still moving. Capacity is not.

That is not a mood. It is the capacity leg of the three-part pattern: demand uncertain, capacity fixed in the short run, labour and material committed before demand is known. This piece stays on that middle line. Not the full fit test again — the bind you already live when the floor cannot invent a second bay by eleven.

What “already committed” means

Committed capacity is not a dashboard tile. It is the set of irreversible or expensive-to-reverse choices already made for the hours left: who is on which shift, which line is open, which van is loaded, which case has cover, which branch can still move stock before close.

A late management report does not reopen those choices. A wall of numbers that never names a measure does not either. Someone still has to decide inside the bind — call in, send home, hold, re-rank, refuse a theatre that pretends capacity can appear from a slide.

If mornings still start by pretending you have free capacity you do not have, the gap is not another chart. The gap is a named proposal against the capacity that is already bound.

Agents propose inside the bind. You keep the switch.

Spectacle is the layer over the operation you already run. Agents on that layer take operational work the owner still does by hand: redeciding, coordinating, assembling what ERP, inventory, and timekeeping already know into what to do next — objects from the operation, not a chat paste.

They propose a named measure against named capacity. You release it. Logged. Reversible where the source system allows. Without approval the agent sleeps — proposal is not release; you keep the switch. Arithmetic stays in code. The model only writes language. We do not write sealed systems of record. We do not auto-assign shifts without human release. We do not invent new capacity out of language.

Capacity theatre is the opposite sell: a pitch that acts as if hours, bays, and cover can be conjured because the model sounds confident. We refuse that. Time back is the want — fewer mornings spent re-ranking a bind that was already set — not invented hours, ROI, euros, or accuracy bands. See also time back, not ROI theatre.

Six bands equal

Bound capacity is not a gastro thesis. The same middle line shows up with different objects. A line that cannot re-tool tonight. A dock that cannot add a bay. A field crew that cannot invent a qualification. A ward that cannot create cover. A branch that cannot move the close. A venue that cannot reprint the shift after the doors open.

Six public bands sit as equals. We name them so nobody reads capacity as one trade’s SKU:

• Manufacturing and production
• Warehousing, logistics and distribution
• Technical service and maintenance
• Health and care
• Retail and multi-site
• Venue operations

Venue ops is last on purpose: equals, not a beachhead story. Pain that spikes when the same bind is copied to a second site is a pattern question — pain at the second site — not a reason to collapse the homepage to one door sign. New sector = configuration on the same core. The layer sits above the systems, not instead of them.

What we are not selling

Not capacity theatre. Not a dashboard that asks you to look harder at a bind you already feel. Not a website. Not a chatbot over bad figures. Not restaurant software. Not “AI for hospitality.” Not a gastro SKU with five other bands as footnotes. Not invented hours, euros, ROI, or accuracy bands. Not silent write-back into the till.

If demand is known far enough ahead, capacity can flex overnight, and you can wait to commit, you already have room. Excel, WhatsApp, and experience may be enough. That is not the buyer this essay is for.

If mornings still start by pretending capacity is free when it is already spoken for, request an initial analysis. Twenty minutes is enough to see whether agents that propose inside your bind — with approval — are worth building. If the data cannot answer the question, we say so.

FAQ

What does “capacity is already committed” mean? Labour, material, dock, machine hours, or shift are already bound for the hours left — before demand has finished clearing. Agents propose inside that limit; they do not invent free capacity.

Is this the same as the three-part pattern essay? No. drei-teil-muster locks the full fit test. This piece deepens the capacity leg only.

Do agents create capacity? No. They assemble a named measure against capacity that is already bound. Approval releases. Without approval the agent sleeps.

Do you invent hours saved by fixing the bind? No. Time back is the want. We do not publish invented hours, ROI, euros, or accuracy bands.

Is this only for hospitality? No. Six equal bands. Configuration, not a beachhead SKU.

Where should I start? Spectacle. The Insights series is how we write about the operation.