Airline fuel efficiency measured on real flights, with the next saving already found
Target: 1% of fleet fuel, about $51 million a year, every gallon measured on real flights.
See one case, screen by screen ↓An airline fuel efficiency program is the set of fuel saving initiatives an airline runs — single-engine taxi-in, APU off at the gate, idle reverse, cost index, water uplift — with the savings each one is expected to deliver. Fuel Efficiency Program is airline fuel efficiency software in which seven agents measure every initiative on real flights, re-check each figure, find new savings with a trial designed and draft the pilot bulletin, while the Director of Fuel Efficiency and the Chief Pilot decide and sign.
How to choose AI for airlines: a buyer's guide to agentic AI solutions →
Updated
Airline fuel efficiency is run as a program of initiatives: single-engine taxi-in, APU off at the gate, idle reverse, cost index optimization by route, potable water uplift, flight-plan performance factors. Each initiative goes into the plan with a claimed saving in gallons a year, and that claim is carried forward into the budget. Proving what each one really saves means pulling FDM/FOQA data, finding the eligible flights, building a matched baseline and working out adoption station by station — about two analyst days per station each quarter, so most initiatives are re-measured once a quarter at best.
Meanwhile the operation moves. Surface metering at LAX cut the median taxi-out from 18.6 to 10.9 minutes, and a delayed engine start that still claims 96,000 gallons a year now measures 11,000. Adoption sits at 23% at one station and nobody can say why. The next saving waits in the flight data until someone has time to look for it, and anything that reaches crews still has to be written, cited and signed by the Chief Pilot.
Estimated split of the two analyst days a quarter it takes to re-measure one initiative at one station by hand, and the review time left with the solution.
Seven agents read every flight, measure each initiative, re-check it, find new savings and draft the bulletin; people approve.
Corvane Airways flies 306 aircraft from its hub at DEN and plans to burn 1.42 billion gallons of fuel in 2026. It is Thursday, October 15. Here is how its Director of Fuel Efficiency works with the agents in the working solution.
Overnight the agents measured all 9 initiatives on 1,402 flights. The program screen is a short ranked list: each line shows its claimed gallons, the gallons measured on flights with a tick for the claim, adoption at the top stations and its status. Claimed 12,586,000 gallons a year; measured 10,703,000, or $38.7M — 0.75% of fleet fuel against a 1% target.
Fuel Program agent: “Morning, Nora. 9 initiatives measured on 1,402 flights from yesterday. One new finding at ORD is waiting for you.”
In 2,812 ORD departures the APU runs a median 47 minutes before pushback on first-wave departures at gates with ground power and PCA; FOM 3.12.2 says 20. The agent sizes it at 197,000 gallons a year, $713,100, showing its arithmetic. The Trial design agent proposes T-038: concourses B and C for 4 weeks, concourse F as control, with success criteria set before the start — a median APU start of 25 minutes or less, no rise in cabin temperature write-ups or delays, ground power and PCA available on 95% of turns.
The Fuel Program agent proposes only; it never starts a trial. Gate C9, whose PCA is still out of service, is left out.
Since bulletin FCB-26-09 went out on June 2, single-engine taxi-in at DEN rose from 41% to 68% of eligible arrivals. Measured on 19,126 arrivals at 8.61 gallons saved each, it saves 312,000 gallons a year — 87% of the 360,000 claimed, $1.13M. The Checker agent re-computed it from raw flight records without the Measurement agent's working: 309,800 gallons, within 1%, with no confounder found.
The agents propose extending it: ORD +97,200 gallons, IAH +70,900 — 168,100 gallons a year, estimated.
IAH sits at 23%. The Measurement agent answers from taxi-time data and de-identified crew feedback: the median taxi-in is 6.4 minutes on the runways nearest Terminal C, and after the 3-minute engine cool-down there are about 3 minutes left on one engine — crews say it isn't worth the workload. On taxi-ins of 9 minutes or more IAH is already at 38%, so the agent suggests a 9-minute threshold at IAH: 52 eligible arrivals a day, target 75%.
Adoption is shown by station and fleet only. Groups under 20 flights are not shown.
T-039 extends single-engine taxi-in to ORD with the 6-minute threshold and IAH with 9 minutes, for 6 weeks from Monday, October 19. Nora picks the reason code TRL-EXT, “Extend a proven initiative to new stations”, ticks that she has read the evidence and signs. The decision goes on the audit trail with her name.
The Bulletin agent drafts flight crew bulletin FCB-26-14 from the operations manual and the measured results: when to use it at ORD and IAH, shut down engine 2 no earlier than 3 minutes after the reversers are stowed (FOM 3.18.4), not on uphill taxi, during de-icing or with an MEL item affecting the APU, hydraulics or the operating engine. Nora sends it to Capt. Daniel Reyes, who approves it with CP-OK, “Within SOP and manufacturer limits”.
“Only the Chief Pilot can approve; nothing reaches crews before he signs.”
The bulletin now carries “Approved and signed by Capt. Daniel Reyes, Chief Pilot · reason CP-OK.” Capt. Lauren Diaz of Flight Standards publishes it to the EFB library, effective Monday, October 19 — no manual change needed. The captain keeps discretion to taxi on two engines, and results are reported by station and fleet only, never by crew member.
Flight 1482 ORD–DEN, Ship 4182, an A321neo on runway 16R: taxi-in 10.4 minutes to on-block, 6.6 minutes on one engine after the 3-minute cool-down, 9.6 gallons saved against the matched baseline. The crew line reads “Not shown: de-identified.” Files for Ships 7307 and 7312 are 21 hours late, so their 6 flights are not counted yet and no saving is overstated.
The Flight data agent flags late or damaged files and never fills a gap by guessing.
The delayed engine start for taxi-out at LAX claims 96,000 gallons a year and measures 11,000. Surface metering since May 4 cut the median taxi-out from 18.6 to 10.9 minutes, the time left on one engine fell to 1.4 minutes and adoption fell from 52% to 34%. The Checker agent tested fleet mix, season and runway works and found none. The agents recommend retiring it as a tracked saving and keeping FOM 3.18.2 as a captain's option; Nora decides with a reason code.
The Reporting agent drafted the Q3 2026 program report: 2,840,000 gallons measured, $10.3M at the finance plan price and 27,192 tonnes of CO2 at the ICAO factor of 3.16. Every figure carries its source and its tag. Nora signs first and Hannah Cole, Director of Sustainability, co-signs.
“CORSIA and EU ETS reports use the fuel actually burned per flight, which already includes these savings. This report explains how much of the reduction came from the program.”
The same fuel efficiency program, seen by five of the people who carry it — what their quarter looked like, and what it looks like now.
Seven agents read every flight, measure each initiative, re-check it, find new savings and draft the bulletin; people approve.
Reads every flight's de-identified FDM/FOQA file and OOOI times each night: engine start and stop, APU on and off, reverser N1, flap setting, fuel flow and taxi times.
Mines all flights every week for new savings, comparing what the fleet does with the manual and the best station and fleet, and sizes each one in gallons and dollars with its arithmetic shown.
Designs a measured trial for each proposal: trial and control groups with the same fleet and equipment, the weeks needed for the sample, and success criteria set before the start, including safety and workload checks.
Measures adoption and savings for each initiative, station and fleet against a matched baseline — same type, station, taxi-time band and outside air temperature — and answers why a number moved.
Re-computes every figure from raw flight records without the Measurement agent's working, and tests confounders such as taxi-time changes, fleet mix, weather and runway works.
Drafts flight crew bulletins from the operations manual and the measured results, quoting each limit exactly, citing it and stating the captain's discretion.
Turns measured gallons into monthly and quarterly cost and CO2 reports, at the finance plan price and the ICAO factor, for flight operations, finance and sustainability.
Approve trials with a reason code; the Chief Pilot signs every bulletin. The agents propose; a named person decides.
Ask about any initiative, station, trial or flight — or tell the agents what to change. Every answer cites its flights or manual passage, and anything that needs a decision waits for one.
Why is IAH adoption low?
Taxi-ins at IAH are short: the median is 6.4 minutes on the runways nearest Terminal C. After the 3-minute engine cool-down that leaves about 3 minutes on one engine, and crews say it isn't worth the workload. On taxi-ins of 9 minutes or more, IAH is at 38% today. I suggest a 9-minute threshold at IAH: 52 eligible arrivals a day, target 75%. It is built into T-039.
Can we tighten the delay-day exception?
Yes, as a proposal. If the higher cost index is used only when 25 or more connecting passengers would misconnect, about a third of the gap comes back: an estimated 240,000 gallons a year ($868,800). Dispatch policy belongs to the Director of Dispatch, so I have changed nothing.
What if the cool-down were 4 minutes?
Each arrival on one engine would lose about a minute of single-engine time: about 1.4 gallons an arrival at DEN, or roughly 51,000 gallons a year less on the 99 arrivals a day that use it. The 3 minutes come from FOM 3.18.4; a manual change is Flight Standards' decision, so I have changed nothing.
Add a rule for winter operations
Added: single-engine taxi is not permitted on contaminated taxiways or below 37°F with visible moisture, citing FOM 3.18.6. ORD will reach that in November, so it belongs in this bulletin. The change is marked in green.
13 screens from the working solution, on its sample data. Pick one to see it large.
How flights are counted, valued and flagged are settings. Changes apply from the next overnight run and are logged.
| Setting | Default | Choose from |
|---|---|---|
| Single-engine taxi-in: eligible taxi-in | 6 minutes | Set in minutes, per FOM 3.18.4 |
| IAH station rule (T-039) | 9 minutes | A longer threshold for short taxi-ins at one station |
| Engine cool-down | 3 minutes | Minutes after reverser stow; time before it never counts |
| Baseline | 8 weeks before the initiative | 8 weeks before the initiative · 12 weeks before · Same weeks last year |
| Fuel price | $3.62 a gallon | Finance plan price, set monthly |
| Below-claim alert | 80% for 4 weeks | Share of the claim measured |
| Retire review | Under 25% for 6 weeks | Share of the claim measured |
| Smallest group shown | 20 flights | Results for fewer flights are hidden |
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 flight data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (IATA — fuel about 31% of 2026 airline operating cost; IATA Fuel Efficiency Gap Analysis — an average 4.4% of fuel savings found per audited airline; ICAO — 3.16 kg of CO2 per kg of fuel; 14 CFR 13.401 and FAA AC 120-82 on FOQA programs and their data). The $51M target is 1% × 1.42 billion gallons × $3.62. Corvane Airways, its people, ships, bulletins and manual sections named on this page are fictional — characters and sample data in the working solution.
An airline fuel efficiency program is the set of fuel saving initiatives an airline runs — single-engine taxi-in, APU off at the gate, idle reverse, cost index, water uplift — with the savings each one is expected to deliver. Fuel Efficiency Program is airline fuel efficiency software in which seven agents measure every initiative on real flights, re-check each figure, find new savings with a trial designed and draft the pilot bulletin, while the Director of Fuel Efficiency and the Chief Pilot decide and sign.
The Measurement agent counts eligible arrivals and the minutes each spent on one engine after the engine cool-down, from the flight data, and multiplies them by the matched two-engine fuel flow — same type, runway, taxi-time band and outside air temperature, in the 8 weeks before the initiative. At DEN that is 8.61 gallons saved per arrival on the sample data. The Checker agent then re-computes it independently from raw records.
APU usage reduction starts with the gap to the manual: the Fuel Program agent compares APU run times with the manual at gates that have ground power and PCA. On the sample data it found the APU running a median 47 minutes before first-wave departures at ORD against 20 in the manual, sized the saving at 197,000 gallons a year, and the Trial design agent proposed a four-week trial with a control concourse.
Not here. The agents work only on de-identified flight data under the agreement with the pilots, and results are shown by station and fleet, never by crew member, with groups under 20 flights hidden. Re-identification stays with the FOQA gatekeeper, for safety, and never with an agent.
Yes. The agents propose, measure and draft; they never start a trial, publish to crews, change the operations manual or retire an initiative. The Director of Fuel Efficiency approves trials with a reason code, the Chief Pilot approves or returns every bulletin with his signature, and Flight Standards publishes.
The Reporting agent converts measured gallons to CO2 at the ICAO factor of 3.16 kg per kg of fuel for internal and board reporting. CORSIA and EU ETS reports use the fuel actually burned on each flight, which already includes these savings; the program report explains how much of the reduction came from the program.
Yes. Eligible taxi-in, station rules, the engine cool-down, the baseline window, the fuel price, the below-claim and retire-review alerts and the smallest group shown are all settings. Changes apply from the next overnight run and are logged on the audit trail.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your program plan, operations manual and the flight data agreement — and it goes live once every quality gate has passed. We will walk you through it on a sample of your own de-identified flight data first.
Recovery plans in minutes, each one rehearsed and costed, signed before anything moves.
1-min demoFlight Dispatch ReleaseEvery release checked before signature, with a source behind every number.
1-min demoNOTAM Briefing238 notices become the 6 that change the flight, each with its reason and original text.
1-min demoWe'll run Fuel Efficiency Program on a sample of your own de-identified flight data and your current initiative list.