Aircraft spare parts demand forecasting that puts the part on the shelf before the aircraft needs it
Parts on the shelf before the aircraft needs them, and idle stock turned back into cash.
See one case, screen by screen ↓Aircraft spare parts demand forecasting predicts how many of each part every station will need, and when, so the part is on the shelf before the aircraft needs it. Aircraft Spare Parts Demand Forecasting does it with eight agents that read component removals, health alerts, the flying, check and fleet plans and the supplier's real delivery record, forecast each part at each station for 30, 60 and 90 days, set min/max stock levels and propose a borrow, transfer, order or sale. The materials planner approves, edits or rejects every one.
How to choose AI for airlines: a buyer's guide to agentic AI solutions →
Updated
Aircraft spare parts demand forecasting at most airlines still means min/max levels set once a year from 24 months of issues in the maintenance and engineering system, reorder points that fire a requisition when stock dips, and the lead time the supplier quoted. The materials planner learns a part is short when maintenance control calls. Brake wear alerts, a slow APU start, a supplier's slipping deliveries and the new aircraft on the fleet plan never reach the forecast.
So a no-go part — one the MEL gives no relief for — grounds an aircraft at one station while the same part sits idle at another. The AOG desk borrows, robs another aircraft or buys at spot, and a spot buy costs about 4.8 times the planned price. Meanwhile slow-moving stock from retiring fleets waits for the annual write-down, and an acquired airline's shelves sit under part numbers nobody has mapped.
Estimated split for one no-go shortfall at a hub station, from the first warning to an approved borrow and order, by hand and with the solution.
Eight agents forecast every part at every station and propose the move; the materials planner approves every order.
Thursday, October 15, 2026, 08:40 at O'Hare. Brake wear alerts are rising on the 787-9s that Corvane Airways bases at ORD. Here is how the agents and the materials planner close the gap, screen by screen, in the working solution.
The agents re-planned every station at 08:12. Marisol's Today screen opens on one card: brake assemblies at ORD run 3 short from Oct 24. Below it, 14 shortfalls ahead at 8 stations, 9 proposals waiting for her worth $1.24M, and $38.6M of idle stock with a proposal. A feed shows what each agent found since 06:00.
“The agents propose to borrow 3 from the pool and order 2 at $96,000 each: $203,700. Decide by 16:00 to catch the pool's 48-hour delivery.”
The stock map shows brake assembly BRK-789-4120, a no-go part, at every Corvane station: ORD 4 on the shelf against 9 needed, SFO 5 / 4, DEN 3 / 2, LHR 2 / 1, NRT 2 / 2. The Meridian pool at DFW has 6 free, and a dashed arc marks the proposed borrow of 3 arriving Oct 17. Beside the map is the conversation with the agents about ORD.
“Label: on the shelf / needed in 30 days · click a station to talk to the agents about it”
Eleven brake positions on six ORD-based 787-9s fell below 20% wear remaining in 9 days, against 4 in a normal 9 days. The first, Ship 2914 brake 3, is at 17% and reaches its 10% removal limit about Oct 24. Every number in the conversation opens its source: here, the aircraft health monitoring brake wear report, position by position.
Confidence 0.93 · 11 alerts read
The part forecast shows 90 days of weekly demand at ORD with its range and the first short day. What moved it, each driver cited: the wear alerts add 3, removals history says 4, flying is up 4% with the added ORD–GRU frequency from Nov 1, and Ships 2945–2948 join ORD from January, so its stock level rises from 6 to 8. Brakes have no MEL relief, so the plan-to level is 98%.
Confidence 0.91 · range 6–10. The forecaster shows a range, never a single number.
Ardent Brake Systems quotes 60 days, but its last four brake orders took 112, 118, 121 and 127 days, so the agents plan on 120: an order placed today arrives about Feb 12, 2027. On Supplier lead times, the watcher also read Halvorsen Hydraulics' email moving PO-48610 from Nov 20 to Jan 8, and proposes 96 days for its pumps — waiting for Marisol to accept.
“Rule: plan on the 80th percentile when actual exceeds the quote by 25%”
Four on the shelf plus 2 back from repair by Oct 29 leaves ORD 3 short from Oct 24. The Sourcing Planner compares a pool borrow of 3 plus an order of 2 ($203,700, covered Oct 17), a transfer from DEN ($200,450, DEN short risk 19%), waiting for the repairs ($192,000, a 61% chance of a grounding) and an AOG spot buy (about $1.38M). The Proposal Checker re-derives the forecast from raw removals and alerts, confirms each pool brake ships with an FAA 8130-3 under clause 4.2, and that $203,700 is inside Marisol's $250,000 limit.
Proposal P-2041 · confidence 0.91 · Proposal Checker: “Agrees · independent check”
Marisol tells the agents what she would try. The Sourcing Planner re-plans: move 1 from DEN by road, here Oct 16 for $650, borrow 2 and order 2 — $200,450, $3,250 less. The Proposal Checker answers in the same thread: DEN would then hold 2 for its own 2 expected changes, with nothing spare until Nov 9, and its chance of a short day rises from 2% to 19%.
Proposal Checker: “I would keep the pool borrow.” · Switch to this plan · Keep the original
Marisol keeps the original and approves P-2041 with the reason “Forecast agreed”. Only then does the Sourcing Planner send loan request LR-58213 to Meridian Component Pool, which acknowledges it, and put purchase order PO-48862 into purchasing for release. ORD turns from short to covered on the map, and each step is on the audit trail.
“ORD covered from Oct 17 · logged”
Decisions lists every borrow, transfer and order the agents propose, sorted by when it must be decided. The Proposal Checker re-derived all nine from raw data: P-2043, bleed air valves for the Halden A220s at BOS, is below 0.80 confidence, and P-2045, $284,000 of engine-driven pumps for SFO, is over the limit, so Angela Brooks-Hale co-signs. Three are under $5,000 and can be approved together.
“Every borrow, transfer and order the agents propose. Nothing is sent until a person approves it.”
Fleet health signals lists the 37 alerts read since 06:00 and the forecast each changed. Eight changed a forecast. One — a brake temperature sensor reading 0 °C on one wheel of Ship 6077 at EWR — looks like a sensor fault, so the Fleet Health Reader did not count it and asked maintenance control to confirm. Maintenance control can correct any signal, and the agents re-forecast.
“Not counted: likely a sensor fault, not wear”
The same spares network, seen by four of the people who carry it — what their week looked like, and what it looks like now.
Eight agents forecast every part at every station and propose the move; the materials planner approves every order.
Turns aircraft health alerts, maintenance messages and wear trends into expected removals per aircraft and date — brake wear, slow APU starts, bleed faults, pump pressure.
Forecasts demand per part, station and horizon — 30, 60 and 90 days — from 36 months of removals, utilization from the flying plan, the check plan, the fleet plan and health-driven removals.
Reads purchase-order acknowledgments, receipts and supplier email, works out each supplier's actual lead time and flags slips on open orders before the part is late.
Sets min/max per station by MEL criticality — no-go, go-if or go — with the service level, the planning lead time and the repair turn.
For each gap, compares a pool borrow, a transfer from another station, an exchange and a purchase on cost, arrival and risk, and writes the proposal with its reason in one sentence.
Re-derives the forecast and the arithmetic from raw data, checks the release certificate and traceability terms of every borrowed or bought part, and the approval limit.
Finds slow-moving and excess stock, matches it to shortages at other stations, and proposes a move, a sale, a return to the pool or a consignment.
Provisions new deliveries against what the airline already holds, and maps an acquired airline's stock to its own part numbers through the parts catalogs' interchangeability.
Approve every order; the director co-signs over $250,000. The agents propose; a named person decides.
Ask a question or give an instruction in plain words. The answer cites its sources and can act — open the right screen, or put a proposal on hold with your reason.
Why does O'Hare need 9 brakes?
Eleven brake wear alerts on six ORD 787-9s in 9 days add about 3 likely changes to the 4 that history predicts; flying is up 4%, and four new 787-9s arrive from January. Brakes are no-go, so the agents plan to 9, likely 7.
Which proposals need the director?
One: P-2045, 2 engine-driven pumps for SFO at $284,000, over the $250,000 limit. Halvorsen's slip to Jan 8 is the reason it is needed now.
How much surplus can we sell this quarter?
$24.1M: A320ceo rotables ($11.2M, best bid 61% of book), A320 engine units ($6.4M), 757 parts ($4.1M) and cabin parts ($2.4M). Another $14.5M can move to stations that need it or go to the pool. Trevor Lindqvist approves each.
Hold every A220 order until the Halden stock is merged
Done: P-2043 (3 bleed air valves for BOS, $184,500) is on hold, with your reason logged. Halden's 2 valves of the older -3301 number may cover part of it once Trevor decides on the one-way items. BOS is short from Jan 14 and the lead time is 88 days, so it must be decided by Oct 20.
13 screens from the working solution, on its sample data. Pick one to see it large.
Aviation inventory optimization here runs on your stock policy: the service levels, lead-time rule and approval limits the agents plan to are settings. Every change is logged, and stock rules and limits change only when the supply chain director approves.
| Setting | Default | Choose from |
|---|---|---|
| Service level, no-go parts | 98% | 99% · 98% · 95% — proposed by the planner, approved by the supply chain director |
| Service level, go-if · go · consumables | 95% · 90% · 85% | Editable; changes need the supply chain director |
| Re-planning | No-go hourly · go-if daily · go and consumables weekly | Set by criticality |
| Plan on actual lead time | On — 80th percentile of the last 6 receipts when it exceeds the quote by 25% | A new supplier lead time is used once the planner accepts it |
| Planner approval limit | $250,000 per proposal | Above it, the supply chain director co-signs |
| Bulk approval | Transfers and pool borrows up to $5,000 | Orders are approved one by one |
| Low confidence | 0.80 | Below it, the Proposal Checker flags the proposal for the planner's judgment |
| Slow-moving threshold | 12 months without movement | Stock a delivery in the next 12 months needs is never proposed for sale |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate (the hours split) · “typical” = published figures (IATA with Oliver Wyman, Oct 2025 — over $11B of supply-chain costs to airlines in 2025, including $1.4B of surplus spare parts holding cost · emergency AOG sourcing at about 4.8 times planned procurement, from published MRO industry data). Rules and forms named: the Minimum Equipment List (14 CFR 121.628) · FAA Form 8130-3 · EASA Form 1. Corvane Airways, Halden Air, Ardent Brake Systems, Halvorsen Hydraulics, Northgate Pneumatics, Meridian Component Pool, the other suppliers and the people named on this page are fictional — characters and sample data in the working solution.
Aircraft spare parts demand forecasting predicts how many of each part every station will need, and when, so the part is on the shelf before the aircraft needs it. Aircraft Spare Parts Demand Forecasting does it with eight agents that read component removals, health alerts, the flying, check and fleet plans and the supplier's real delivery record, forecast each part at each station for 30, 60 and 90 days, set min/max stock levels and propose a borrow, transfer, order or sale. The materials planner approves, edits or rejects every one.
In most MRO inventory planning, from removal history in the maintenance system, with min/max levels reviewed once a year. Here the Demand Forecaster uses 36 months of removals and adds what history cannot see: health-driven removals from the Fleet Health Reader, utilization from the flying plan, the check plan and new deliveries. Each forecast is a range with its drivers cited — in the demo, ORD's brakes go from 4 on history alone to 9 at the 98% level.
The Stock Level Setter reads each fleet's Minimum Equipment List (the MEL, under 14 CFR 121.628 for US carriers) and marks every part no-go, go-if or go, then sets min/max per station from the service level for that criticality — 98%, 95%, 90% and 85% for consumables by default — the planning lead time and the repair turn. It only proposes; a level changes after approval.
It aims to see the shortfall early enough to borrow or move a part. The design goal is 40% fewer aircraft grounded for a part. In the working solution's back-test on 24 months of sample data, 31 of 44 parts-caused ground events would have been seen at least 10 days ahead; 5 had no warning in any data, and the results screen grades each miss.
No — it uses what the supplier actually delivers. The Lead Time Watcher reads acknowledgments, receipts and supplier email, and when actual lead time exceeds the quote by 25% it plans on the 80th percentile of the last 6 receipts. A new planning lead time is used once the planner accepts it, and slips on open orders re-plan the shortfalls they affect.
The Surplus Finder lists stock with no movement in 12 months or more and proposes to sell it, move it to a station whose forecast needs it, consign it or return it to the pool. It keeps anything new deliveries will need and every insurance spare. Nothing is sold, moved or consigned without the inventory manager.
Yes. The agents never send a loan request, purchase order or supplier message on their own. A named planner approves, edits or rejects each proposal with a reason, the supply chain director co-signs above $250,000, and no borrowed or bought part goes ahead without an FAA 8130-3 or EASA Form 1 release certificate.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your stock policy, MEL criticality, approval limits and a sample of removals, stock and supplier history — and it goes live once every quality gate has passed. We will walk you through it on your own stations first.
Every caterer line checked against what was loaded, and every missed penalty claimed.
1-min demoGalley PlanningEvery galley loaded to spec: no cabin short of a meal, no weight flown for nothing.
1-min demoCatering Demand ForecastingMeals ordered for the passengers who board, not the ones booked, and no cabin left short.
1-min demoWe'll run Aircraft Spare Parts Demand Forecasting on your own removals, stock and supplier history.