The game went long. Wrap slipped two hours. Twenty-two people in that compound have somewhere to be tomorrow, in eleven different cities. Apron watches the show, the itinerary, and everything that impacts whether the plan holds, and keeps every next call intact.
Original booking is a solved problem. Your TMC does it well, weeks out, against a schedule everyone agreed to. The failure isn't in the booking. It's in the four hours after a live event decides to run long — and the person holding that bag is standing in a truck compound with a phone and a spreadsheet.
Your A1 misses the Monday call. You backfill with a local who has never seen the show, never mixed that booth, and doesn't know the director's shorthand. The position is filled. The show is not the same.
IATSE and NABET-CWA measure turnaround wrap-to-call. Travel time counts as work, not rest. A generic travel tool will happily book a 06:05 flight that satisfies nobody's rest clock — and the grievance lands three weeks after the show.
Twenty-two crew, eleven destinations, one after-hours desk, and a coordinator who has been awake since the 06:00 truck call trying to remember who has an early Monday and who is just going home.
This is what makes live crew travel unlike corporate travel or team travel. The team charters. The crew disperses. Most are going home — to eleven different home markets — and a handful are going straight to the next show, sometimes for a network that isn't yours. There is no single group itinerary to fix. There are twenty-two individual ones.
This has always been a hard night. Three things are making it a more frequent one — and the fourth is the baseline everyone is planning against without saying so out loud.
The question was never whether someone on tonight's crew gets disrupted. It's how many, which positions, and whether anyone catches it while there is still time to do something about it. That third one is the whole product.
Three jobs, run continuously from first pitch until the last A-list position is wheels-up. Not a booking tool — an irregular operations desk that happens to be pointed at your rundown.
Score, inning, clock, weather, rundown, plus flight status and airport ops. When the projected final moves, the wrap-to-gate chain recomputes for every traveling crew member — before anyone has missed anything.
When slack starts to die, Apron generates real options for each individual against their own constraints: their home market, their next call, their rest clock, their union's rules, their gear. Ranked by whether they actually make it.
Most tools have no concept of finished. Apron does: every A-list position is confirmed en route to a call it can make, with rest intact. Until then the board stays open — and it tells you exactly who is still outstanding.
Not one model with a big prompt. Each agent owns one domain, holds one credential, and reads only what its job requires. Crew is provisioned per show: a domestic regular-season game doesn't call CUSTOMS to work.
An agent that reads live external feeds is an agent that can be fed something hostile. An agent that can book travel is an agent that can spend your money. Apron never makes them the same agent.
SPOTTER and TRAFFIC ingest untrusted third-party data — scoreboard feeds, weather services, carrier status APIs. They have zero write authority and zero access to personal data. Anything arriving in those feeds is treated as an observation to be evaluated, never as an instruction to be followed.
FIXER searches inventory and places holds against tokenized travelers — a loyalty tier, a home market, a rest window, a gear profile. It does not receive names, contact details, employment records, or payment instruments. The re-identification happens on the desk side, behind the human gate.
Agents hold; people ticket. Every action that spends money, cancels inventory, or reaches a crew member passes through a named human on the travel desk with the full option set and rationale in front of them. The autonomy is in the analysis, not the authority.
| Agent | Reads | Authority | Never receives |
|---|---|---|---|
| SPOTTER | Game feeds, clock, weather, rundown | READ ONLY | Crew identity · PNRs · any personal data |
| TRAFFIC | Carrier status, ground ops, airport conditions | READ ONLY | Crew identity · next calls · fare data |
| ADVANCE | Crew sheets, call sheets, roster provenance | ROSTER STATE | Payment instruments · fare detail · HR records |
| STEWARD | Agreement rule set, wrap and call timestamps | RULE FLAGS | Contact info · payment · personnel files |
| WRANGLER | Constraints crew choose to disclose | CONSTRAINTS | Undisclosed next-call detail · third-party call sheets |
| FIXER | Inventory, fare rules, travel policy, tokenized traveler | HOLD ONLY | Names · contact details · payment instruments |
| RUNNER | The approved decision record | HANDOFF | No outbound channel to crew, ever |
| CUSTOMS | Carnet requirements, permits, border rules | DOC FLAGS | Fare data · payment · financial records |
Each fact on the board records which agent produced it, which source it came from, how confident it is, and who is allowed to see it. An inferred next call looks visibly different from a crew-confirmed one, and the system will not act on the inference.
Every agent observation, option generated, hold placed, rule flag raised and human approval is timestamped and attributable. Export the night as a single record for your own review, your TMC's, or a labor inquiry.
Freelance crew work for a lot of people. The camera op in your compound tonight may be on a competitor's show Monday. That's their business, and their next employer's — not yours, and not ours. Apron is built around that boundary instead of around it.
Your coordinator gets the production seat — their show, their crew, their night. Your travel management company gets the desk: every show on the board, every network, every PNR under watch. The same incident, rendered for the person who has to act on it.
Every monitored PNR is a line item with a state, a watch window, and a reason it's being watched. Roll it up by production, by network, by client. The commercial model matches how a TMC already meters work — no savings-claim arithmetic required.
The number the desk manager actually wants: how many disruptions the agent crew resolved to a decided plan versus how many rang an agent's phone at 2:00 AM. Broken out by show, by night, and by reason code.
Escalations arrive pre-worked, not raw — options generated, policy checked, rule conflicts flagged, provenance attached. The human decision is approve or redirect, not investigate from zero.
Preferred carriers, negotiated fares, class-of-service rules and approval thresholds come from the TMC's existing policy engine. Apron proposes inside those walls — it doesn't route around them to find a cheaper seat.
A TMC serving four networks runs four sealed tenants. No cross-client inference, no shared roster memory, no pooled model context. A freelancer appearing on two clients' shows is two unlinked records.
The coordinator raises an issue from the compound; the desk sees it with full detail. The desk resolves it; the coordinator sees the outcome with external commitments redacted. One incident, two disclosure levels, no re-keying.
Your coordinator is in a truck compound with a headset on and a director in their ear. A dashboard only helps the person looking at it. Apron Alerts is how the board finds them when they're not.
Not by how loud it is. Watch means we're tracking it and you don't need to know yet. Decide means a human is required. Clear means it's finished. Only one of the three interrupts anything.
Apron fires when a decision is needed, not when something changed. A coordinator who gets forty pings a night ignores the forty-first — which is always the one that mattered. Restraint is the feature.
The production coordinator gets their show. The TMC desk gets the escalation with the full option set. Same incident, two alerts, two levels of detail.
An external call stays withheld in the alert exactly as it is on the board. The redaction is enforced at the data layer, not the display layer — so it survives every channel.
Which agent raised it, which source it came from, how confident it is, and whether it's confirmed or inferred. Enough to act on without opening a laptop to check whether it's real.
SMS and push for the compound. Straight into the TMC's existing exception queue for the desk. No new app for a coordinator to install at 1:00 AM on the worst night of their season.
Apron sits on top of the travel program you already have. Your TMC keeps the booking, the contracts, the servicing relationship and the ticketing authority. We take the four hours they were never staffed to cover.
Mid-office feed, GDS or NDC queue, or your TMC's own API. If your TMC is already on Apron, the production seat comes with the program — nothing for the broadcaster to procure.
Crew sheets and call sheets in, plus live game state. Works from a spreadsheet on night one; integrates with your scheduling and crewing system when you're ready.
The applicable turnaround, rest and travel-time provisions are configured per production and per local, so the rule check reflects the deal your show actually works under.
It does not book your original travel — that's well covered and we're not trying to win it. It does not staff or crew your show. It does not message your crew directly. It does not hold your payment instruments. It does not claim a savings number, because the value here is a covered position at 14:00 Monday, not a cheaper seat.
The watch board plays a real irregular-operations night end to end — from "projected final revised" in the bottom of the ninth to the last A-list position confirmed en route. Switch between the production seat and the TMC desk to watch the same incident change shape depending on who's allowed to see what.
Not per seat, not per user, not a platform fee. You pay for the itineraries actually under watch, for the window they're under watch.
$0.50 at volume. Billed through your travel management company alongside the transaction fees you already pay — one line, one relationship.