Airline network planning decides what an airline flies: which cities, how often, on which aircraft, and at what time of day. Every new route takes an aircraft from somewhere else, so every route has to earn its place.
This guide is for anyone who plans, approves or pays for an airline's network: network planners and route analysts, schedule and fleet planners, slot and alliance teams, and the commercial leaders deciding whether to automate the work. No background needed.
Imagine you plan routes for an airline with a hub in Chicago. You want to fly to Athens four times a week next summer on a Boeing 787-9, starting in early June. The forecast looks good: 82% of seats filled.
But from June 3 to June 28, one of your 787-9s is in the hangar for a heavy maintenance check. To fly Athens in June, you would have to use your only spare aircraft.
Your route idea
Route
Chicago–Athens
Frequency
4 a week
Aircraft
787-9
Starts
June 3
The fleet calendar
N812CV
heavy check
Out
June 3–28
Spare
needed for Athens
From July 1
a 787-9 is free
June: one 787-9 shortJuly 1: it fits
Find that after the route is on sale, and you lease an aircraft with crew, start late, or change flights for passengers who already hold tickets.
Good network planning finds the missing aircraft before the route goes on sale.
The seasons, the new-route study, hub banks, fleet, slots and the rules, explained without selling anything. If you are new to network planning, start here.
01 · In plain words
Airline network planning in five minutes
An airline's network is the list of routes it flies: which cities, how often, on which aircraft and at what time. Network planning decides that list, one season at a time. It is the airline's largest commercial lever, because pricing and revenue management can only sell the seats the network puts in the sky.
To add a route, planners ask a chain of questions. How many people travel between the two cities? What share would we win? How many passengers would connect from our other flights at the hub? Do we have an aircraft every week? Can we get takeoff and landing times at both airports? What does it cost, and what does it take from our other flights?
Each answer comes from a different team. So a new-route study is a relay: it passes from desk to desk, and most of the time goes to waiting for the next hand-off.
The life of one route idea, in five steps
An idea"Chicago–Athens, four a week, next summer, on a 787-9."
DemandForecasting sizes the market: about 82% of seats filled.
Feed and fleetConnections at the hub, and an aircraft every week. This is where the June gap is found, or missed.
Slots, partners, cost, decisionThe slot team, alliances and finance add their pieces; a senior leader signs.
On saleMajor US carriers open sales about 11 months ahead. A change after that means rebooking ticketed passengers.
Step 3 decides this example. Miss it, and step 5 finds it for you.20
Four things that surprise people
The best US data misses part of the market
The US ticket survey covers tickets issued by US airlines. Foreign airlines do not report, so a transatlantic market needs other sources too.2
Forecasts often assume rivals stand still
The demand forecast usually assumes no competitor responds. The competition view lives in another team's deck, so the downside is easy to miss.
Growth means moving aircraft, not buying them
The industry had 18,100 aircraft on order in May 2026, and North American capacity is forecast to grow just 0.3% in 2026. A new route takes a tail from somewhere.1
A takeoff time you don't use, you lose
At busy airports, a series of slots keeps its priority next season only if it was flown at least 80% of the time.4
Four words you will see on this page
O&D
Origin and destination: where a trip really starts and ends, whatever the connections. Markets are sized this way.
Hub bank
A wave of arrivals followed by a wave of departures, timed so passengers can connect.
Heavy check
Major scheduled maintenance that takes an aircraft out of service for weeks.
Slot
Permission to land or take off at a busy airport at a set time.
In shortEvery route is five decisions: where, how often, on what aircraft, at what time, and in which season. Network planning makes them between fleet planning, years ahead, and pricing, months ahead.
Airline network planning decides what the airline flies, season by season. It is the largest commercial lever an airline has: pricing and revenue management can sell the seats the network puts in the sky, and no others.
Five decisions in every route
WhereCity pairWhich origin and destination to connect, and through which hub.
How oftenFrequencyDaily, four times weekly, seasonal: each level changes share and cost.
On whatGaugeThe aircraft type and seat count; range and payload set the limits.
WhenTime and bankThe departure time, and which wave of the hub day it joins.
Which seasonSeasonIATA summer or winter, start and end dates, and the ramp-up.
The classic airline planning sequence
Airline planning is taught as a sequence that runs from years out to the day of operation 16. Network planning sits in the middle: it takes the fleet it is given and decides how to deploy it.
1 · YearsFleet planningHow many aircraft of which types, and when they arrive or retire.
2 · SeasonsRoute evaluationWhich markets to serve, how often, and with what network structure.
3 · MonthsSchedule developmentTimetable, banks, fleet assignment and rotations.
4 · Months to daysPricingFares by market, filed for sale.
5 · DaysRevenue managementWhich fares to sell on which flight, day by day.
Network planning owns steps 2 and 3 together with schedule planning. This guide focuses on them.
Who decides
Owner
SVP Network Planning
Signs each version of the season plan and answers for network revenue and margin.
Budget
Chief Commercial Officer
Holds the commercial budget; network, pricing and revenue management usually report here.
Daily work
Network planners and analysts
International and domestic desks frame route ideas, run studies and recommend versions.
03 · Seasons and the calendar
The airline schedule planning calendar: from network plan to publication
In shortA route decided today flies about a year from now. Plans are built by IATA season, and each season has a fixed slot calendar the plan must meet.
A route decided today flies about a year from now. Network plans are built by IATA season, and each season has a fixed slot calendar that the plan must meet.
The planning horizon (not to scale)
Swipe sideways to see the full diagram.
Schedule development starts a year or more ahead, and timetables and rotations are set two to six months out 16. Major US carriers open sales about 11 months before departure 20. Moving a route after it is on sale means rebooking passengers who already hold tickets.
IATA seasons
Northern summer
Last Sunday of March to last Saturday of October
Summer 2027 (S27) runs from March 28 to October 30, 2027 6.
Northern winter
Last Sunday of October to last Saturday of March
Winter 2026/27 (W26/27) began on October 25, 2026.
The slot calendar for summer 2027 As of October 2026
Coordinated airports allocate slots on a worldwide calendar 5, and the FAA sets its own schedule deadline for US Level 2 and Level 3 airports 6. A route idea that misses the initial deadline is handled after the initial allocation.
SEP 14, 2026Historic slots listCoordinators publish the slots each airline holds by history.
OCT 8, 2026Initial submission deadlineAirlines submit their summer 2027 slot requests.
OCT 16, 2026FAA schedule deadlineS27 schedules due for ORD, JFK, EWR, LAX and SFO.
NOV 5, 2026 · NEXTSlot allocation listCoordinators return initial allocations.
NOV 17–19, 2026IATA Slot ConferenceAirlines and coordinators trade and adjust slots.
DEC 15, 2026Series return deadlineReturn unwanted series without losing future priority (trial date).
JAN 31, 2027Historics baseline dateThe usage count toward next summer's history starts here.
MAR 28, 2027Summer 2027 startsThe season runs to October 30, 2027.
Dates from the IATA coordination calendar and the FAA notice of October 9, 2026 56. Check each airport's coordinator for local deadlines.
04 · The new-route study
A new-route study is a relay between teams
In shortNo single team can answer "should we fly this?" alone. The study passes from desk to desk, and most of the elapsed time is waiting for the next hand-off.
No single team can answer "should we fly this?" alone. The study passes from desk to desk, and each one adds a piece. Most of the elapsed time is waiting for the next hand-off.
1 · Idea
Frame itNetwork planning sets route, frequency, aircraft, season and time.
2 · Demand
Size itForecasting sizes the O&D market and share; pricing adds the fare view.
3 · Feed
Connect itSchedule planning places it in a bank and counts connecting feed.
4 · Fleet
Find a tailFleet and maintenance planning check aircraft, heavy checks and spares.
5 · Slots and partners
Clear itThe slot team, alliances and government affairs check slots, rights and partner flows.
6 · Decision
Price and signFinance prices it; the SVP Network Planning decides.
Who does what
Team
Its piece of the study
Network planning
Frames the idea (route, frequency, aircraft, season, time) and owns the recommendation.
Demand forecasting and route analysis
Sizes the O&D market and the airline's share; forecasts passengers, load factor and revenue by month.
Pricing and revenue management
Fare and yield view; what the new flight displaces from other traffic.
Schedule planning
Departure time, bank placement, connections and rotations.
Fleet and maintenance planning
Is there a tail? Heavy-check calendar, operational spares, range, payload and ETOPS.
Slots and airport affairs
Slot or schedule approval at coordinated and facilitated airports; curfews; terminal and customs capacity.
Alliances and partnerships
Partner feed beyond the destination; codeshare and joint-venture terms; DOT filings.
Finance
The cost model (fuel, crew, maintenance, fees, ownership) and contribution.
Government affairs and legal
Traffic rights under air services agreements; route authority.
SVP Network Planning, then the CCO
Decision, and sign-off of the season plan version.
Where the relay hurts
Every hand-off waits
Each team works its queue. The study moves at the pace of the slowest desk, so few versions get tested.
Costs: one version reaches the SVP instead of three
Disagreements stay hidden
The demand forecast often assumes no rival response, while the competition view lives in another deck.
Costs: a downside nobody priced
Fleet limits arrive late
A route is promoted, then fleet finds no spare aircraft in the month a heavy check takes a tail out.
Costs: a wet lease, a late start, or changes after sale
Assumptions can't be traced
Six months later, nobody can say which data month, fuel curve or feed assumption drove the decision.
Costs: decisions re-argued from scratch
How long does a study take? No published industry benchmark exists for new-route study time or effort. Measure your own: count the working days from idea to recommendation on your last few studies, and where each one waited (section 20, question E1).
05 · Evaluating a new route
How airlines evaluate a new route: O&D demand data and route profitability
In shortStart with how many people travel, then the share you would win, the passengers your hub adds, what the route takes from your other flights, and what it costs to fly. Then check whether it can fly at all.
New route evaluation moves from the market to the network: how many people travel, what share the airline wins, what the hub adds, what the route takes from the airline's other flights, and what it costs to fly.
1O&D demandSize the origin-and-destination market, not just the nonstop segment.
2ShareEstimate the airline's share against competing itineraries with a quality of service model 22.
3SeasonalityShape the forecast by month: shoulder months are softer than the peak.
4Connection feedAdd the passengers who connect into the flight from spoke cities.
5Spill and recaptureAllow for demand turned away, and for traffic the route takes from the airline's own flights 17.
6Rival responseShow the forecast if a competitor matches, as a probability with its basis.
7Route profitabilityPrice it line by line to route contribution, then network contribution.
8ConstraintsCheck fleet, slots, traffic rights, connect times, curfews and ETOPS.
O&D demand data: what planners use in the US
DOT O&D surveySince July 1, 2025, a 40% monthly sample of tickets, reported by the US carrier that issued them 2, published as DB1C 3. The older DB1B series was a 10% quarterly sample.
T-100 trafficMonthly segment and market traffic from US carriers and foreign carriers serving the US. International data have historically lagged by about six months 23.
Your own bookingsBooking and one-stop history in the market, fares and fare history.
Booking-channel dataGDS booking data (MIDT) and settlement-based data; GDS data miss direct web, app and call-center sales.
Partner dataPartner sales and onward flows beyond the destination.
Published schedulesEvery carrier's schedule loads: who flies the market today, and who just added it.
Fare shoppingCompeting one-stop fares, which set the price a new nonstop must beat.
Finance cost modelFuel curve, block-hour rates, fees and ownership, with an as-of date.
The gap in the survey. The DOT survey covers tickets issued by US carriers. Foreign carriers do not report 2, so a transatlantic or transpacific market needs other sources to count tickets that foreign airlines sold.
Airline route profitability: from revenue to network contribution
Worked example · Chicago (ORD) to Athens (ATH), 4x weekly on a 787-9, July 1 to October 30 (sample data from a fictional airline)
Swipe sideways to see the full diagram.
Cost lines, left to right: fuel $10.6M, crew $3.4M, maintenance $2.6M, airport and navigation fees $3.0M, ownership $3.9M, handling, catering and distribution $3.9M. "Taken from own flights" is the revenue the new route draws from the airline's existing itineraries over other gateways. Sample data from a fictional airline, Corvane Airways; not a forecast for any real market.
06 · Hub banks and feed
Hub banks, minimum connecting times and connection feed
In shortA long-haul flight from a hub is filled by local passengers and by passengers connecting from other cities. The time of day it leaves decides which of those connections can make it.
In a hub-and-spoke airline, a long-haul flight is filled by two groups: local passengers, and passengers connecting from spoke cities. The hub bank structure decides which spoke cities can reach a departure in time.
One bank at a hub: arrivals land, passengers connect, departures leave
Swipe sideways to see the full diagram.
A departure at the end of a bank can be fed by every spoke arrival that lands at least the minimum connecting time before it. Move it one bank earlier and late arrivals miss it; move it later and some spoke cities have no flight that lands in that bank.
Feed by departure bank (sample data)
The same long-haul flight placed in three banks at the fictional airline's Chicago hub. In the sample, 46% of the flight's passengers connect.
Minimum connecting time: the types that matter
Domestic to domestic
Same terminal or a short walk; the shortest connection.
Domestic to international
Adds check-in rules and terminal moves.
International to domestic
In the US, includes customs and border clearance; usually the longest.
MCTs are set by airport, terminal and carrier. Use your published values, not a rule of thumb.
Beyond feed counts too. Passengers who continue past the destination on a partner's flights add value to the route, and a codeshare with a foreign partner needs a DOT statement of authorization 21.
07 · Fleet assignment
Fleet assignment and the heavy-check calendar
In shortEvery new route takes an aircraft from somewhere. Maintenance takes aircraft out for weeks at a time, so a route can fit in one month and be short in the next.
Fleet assignment matches aircraft types to flights to earn the most contribution; tail assignment later puts specific aircraft on lines of flying around their maintenance. A new long-haul route always takes an aircraft from somewhere.
Worked example · one widebody fleet, summer 2027 (sample data from a fictional airline)
Swipe sideways to see the full diagram.
With one aircraft in heavy check, a June start would take the operational spare. Three fixes, from cheapest: start on July 1; borrow an aircraft planned for another route's up-gauge; or wet-lease a widebody with crew for the gap.
Range and payload
Can the aircraft make it?
Range depends on seats and cargo; on long sectors payload falls. Twin-engine flights more than 60 minutes from an adequate airport need FAA ETOPS approval 13.
Deliveries
New aircraft arrive late
The industry backlog stood at 18,100 aircraft in May 2026 1. Growth comes from moving the tails you already have.
Crew and bases
Aircraft need crews
Pilot qualifications by type, crew bases and pairings shape which routes a fleet can fly.
08 · Slots, rights and rules
Airport slot coordination, the 80/20 rule and traffic rights
In shortA route can win on demand and fleet and still not fly. Busy airports ration takeoff and landing times, international routes need traffic rights, and changes after sale owe passengers refunds.
A route can win on demand and fleet and still not fly. Slots, hub limits, traffic rights and consumer rules all sit on the route's path. Statements in this section are As of October 2026.
Worldwide Airport Slot Guidelines, Edition 4, effective August 1, 2025, published jointly by airports, airlines and coordinators.
80/20 use-it-or-lose-it: a slot series keeps historic precedence next equivalent season if it was flown at least 80% of the time.
Summer 2027 minimum series: a summer series must hold at least 7 slots, up from 5. Winter stays at 5.
The 80% threshold remains debated; IATA argues raising it would not raise peak use 18.
US airports: FAA slots and limits
Airport
Status, as of October 2026
Source
JFK
Level 3 under an FAA order. Carriers must use each slot at least 80% of the time; a staffing waiver lets carriers return up to 10% of slots through October 30, 2027.
Level 2. An FAA order caps scheduled operations at 2,708 a day from 06:00 to 23:59, with half-hour limits of 30 to 84, through October 30, 2027. New requests that would exceed a limit are not accommodated.
The FAA notes that approval for runway access is separate from approval for terminal space 6. Abroad, coordination is tightening too: press reports in 2026 say Athens moves from Level 2 to Level 3 for summer 2027 19confirm with the coordinator.
Traffic rights, consumer rules and the rest
Traffic rights
Air services agreements
International service needs rights under the agreement between the two countries. Open-skies agreements remove limits on routes, capacity and pricing. Under the 2007 US–EU agreement, US carriers may fly from the US to any point in any EU member state 12. Restrictive bilaterals still limit some markets.
Partners
Codeshare authorization
A codeshare with a foreign partner needs a DOT statement of authorization under 14 CFR Part 212 21. Check current filing lead times with counsel.
Consumer rules
Changes after sale
Under 14 CFR Part 260, in force since October 28, 2024, a passenger whose flight is cancelled or significantly changed (for example, 3+ hours domestic or 6+ hours international, a different airport, or more connections) and who does not accept the alternative is owed a prompt refund 1011.
DOT is not enforcing refund duties for flights that are merely renumbered with no significant change, until July 7, 2027 11. Other constraints a study checks: curfews, terminal space, customs-hall capacity, crew bases and route authority.
09 · Why it matters now
Every aircraft has to earn its route
In shortMargins are thin, fuel is expensive and new aircraft arrive late. Growth comes from moving the aircraft you already have, so every route decision counts.
Six numbers that frame network planning in late 2026, each with its source.
2.0%
industry net margin forecast for 2026: $23.0B of profit, down from $45B in 2025.
The DOT O&D survey quadrupled its sample and went monthly in July 2025: fresher data, and more of it to reconcile 2.
A summer slot series needs at least 7 slots from summer 2027 4.
FAA limits at ORD and EWR, and New York slot relief, run through October 30, 2027 789.
Peer-reviewed research finds new air services are often abandoned within their early years, and that hub connectivity improves survival 15.
A note on the numbers. No published industry figure exists for route-launch failure rates, new-route study time, or the cost of a wrong route. Part C gives you questions and a do-it-yourself calculator instead (sections 17 and 20).
10 · How teams do it today
Three ways airlines run a new-route study
In shortMost airlines combine the first two. The real question is how many versions of a route reach the decision, and whether fleet, rivals and slots are checked before the route goes on sale.
1 · The relay
Decks and spreadsheets, team by team
The network desk frames the idea and passes it to forecasting, schedule, fleet, slots, alliances and finance in turn. Each team adds its piece.
Catches
What each expert knows about their own piece
Misses
Conflicts between pieces, late fleet limits, and which assumption drove the decision
2 · A planning suite
A network and schedule optimization engine
Software that is fast at the arithmetic: market sizing, share, schedules and scenarios.
Catches
The math, quickly, on many scenarios
Misses
The cross-team study around it: fleet calendar, slots, partners and cost still pass by hand-off
3 · The whole study
Software that answers every question at once
Newer tools, some built on AI, aim to answer demand, feed, fleet, rivals, slots, partners and cost for the same what-if, with sources.
Catches
Conflicts and constraints before a version is chosen
Watch for
Whether every number is cited, and who is allowed to promote and sign
The relay and the whole study, side by side
Question
The relay between teams
The whole study at once
How long?
At the pace of the slowest desk
The same sitting, for every version
How many versions reach the decision?
Often one
Several, side by side
Rival response?
In another team's deck, often left out
A matched case on every forecast, with its basis
Is there an aircraft?
Checked late, sometimes after the route is on sale
Checked against the heavy-check calendar before a version is chosen
Can you trace the decision later?
Hard: which data month, fuel curve or feed assumption?
Every number with its source and as-of date
Whichever way you work, people make the decision: a planner recommends, the fleet team confirms, and a senior leader signs the season plan.
Part B · See it working
Botminds Airline Network Planning, one route at a time
Part A described the work without naming a product. This part shows how Botminds does it, at a fictional airline, Corvane Airways, with its hub at Chicago O'Hare. Every name, tail number and figure in the screens is sample data.
11 · One route, start to finish
Follow the Chicago–Athens idea through the study
In shortThis is the example from the top of the page, inside the product. Eight agents answer the same what-if at once. People confirm, promote and sign.
1
Ask once, on the route map
The planner types "ORD–ATH four times weekly next summer on a 787-9". A dashed arc draws from Chicago to Athens, and the agents pin their answers to it: demand $38.4M at an 82% load, 23 cities feeding through the hub, contribution $5.1M, $2.3M of partner traffic beyond Athens, slots to request, and a rival that may match.
One pin is red. Fleet: June short.
Network map. Every agent's answer sits on the route it is about, with numbered sources in the thread beside it.
2
The missing aircraft, found first
From June 3 to June 28, N812CV is in heavy check, so Athens would take the operational spare. The Fleet agent offers three fixes, cheapest first: start on July 1; use the aircraft planned for Dublin in June, about $0.4M; or lease a widebody with crew, about $1.6M.
The agent proposes. Tom Reyes, the fleet planning manager, confirms.
Fleet fit. Tails by week against lines of flying, the June heavy check, and three fixes from cheapest.
3
The rival priced in, in the open
Demand says 82% with no rival response. Competition says 76% if Northmere matches. The two answers are shown side by side, each with its sources.
The Review agent names what would settle it: Northmere's summer 2027 schedule load, expected in January. Nobody hides the downside.
Disagreement in the open. Both numbers are cited, with the evidence that would settle them and when it is expected.
4
A person promotes it, with a reason
Version B starts on July 1: $33.6M of revenue at an 85% load, contribution $6.2M, and the fleet fits. Elena Brandt, a senior network planner, promotes it with a reason code. The open disagreement goes into the brief with both sources.
Agents never promote, sign or file. The SVP Network Planning signs the season plan.
Promote with a reason. A reason code, an attestation to the sources, and the open disagreement carried into the brief.
12 · Every screen
The rest of the study: versions, the rival, hub banks, sign-off and the audit trail
In shortOne route is the story. These screens show the rest: changing one thing and re-running only what it touches, comparing versions, the hub bank view, the leader's sign-off, and who may do what.
Change one thing. Move the start to July 1 and four of eight agents re-run; the others' answers still hold. Contribution rises and the fleet fits.
Ask back. "What if Northmere matches?" The Competition agent answers with the effect on load and contribution, and states its basis as a pattern, not an announcement.
Ask in plain words. "Why did the Fleet agent flag ORD–ATH?" The answer cites the heavy-check record and opens the fleet screen.
Compare versions. Three versions of the same route side by side, with the best value in each row marked and the reasons they differ. Sample data, fictional airline.
Hub banks. Eight banks a day at the hub, the long-haul departure at the end of bank B6, and the feed it collects. Choosing B5 or B7 recounts the feed. Sample data.
Sign-off. The SVP Network Planning signs the season plan version, with every change since the last version and the Review agent's brief. Exceptions past their service level are escalated.
Audit trail. Every agent run and human decision as it happened, kept unedited, with an evidence pack to export.
Roles. Who may promote, confirm, approve and sign. The agents' row is locked: they compute, build versions and draft, and never promote, sign, confirm or file.
Select any screen to open it full size. Every name, tail number and figure on these screens is fictional sample data.
13 · Agents, and who signs
Eight agents answer the study. Four people make the decisions.
In shortEach agent has one job and one thing it never does. Agents compute, build versions and draft; people promote, confirm, approve and sign.
The eight agents
Demand
Sizes the O&D market and the airline's share; forecasts passengers, load factor and revenue by month.
Never assumes a rival response, or forecasts on stale data without saying so.
Connection
Finds the spoke cities that feed the flight through its bank, and the connecting share.
Never moves another flight to make a connection work.
Fleet
Checks tails by week against lines of flying and the heavy-check calendar; range and payload.
Never assigns or swaps a tail; it proposes fixes for the fleet manager.
Competition
Reads rivals' schedule loads and launch history; estimates a response and its effect.
Never states a rival's intent as fact: a probability with its basis.
Slots and rights
Checks slot coordination, curfews, traffic rights and the customs hall; drafts the slot request.
Never files with a coordinator.
Partner
Counts partner flows beyond the destination, values a codeshare, and drafts the DOT request.
Never sends a filing or commits the airline to a partner.
Economics
Prices the route line by line: contribution after what it takes from the airline's own flights.
Never changes a cost assumption without the planner.
Review
Re-derives revenue and cost blind, names disagreements and what would settle them, and writes the brief.
Never promotes, or resolves a disagreement itself.
People decide; everything is cited
PromotesNetwork plannerMoves a version into the season plan, with a reason code and an attestation.
ConfirmsFleet planning managerConfirms the fleet finding or offers a tail, with a reason.
ApprovesAlliance lead and slot teamReview the drafted codeshare and slot requests, and file them.
SignsSVP Network PlanningSigns the season plan version, or sends it back with a reason.
Every number carries a source with its as-of date. Stale feeds are flagged before anyone relies on a number, and the answers that depend on them are marked. Promoted changes go to scheduling as SSIM. Every agent run and decision lands on an audit trail that cannot be edited.
Design targets, to be measured in a pilot
0
routes promoted without an aircraft to fly them: every what-if is checked against the heavy-check calendar before promotion.
Design target
3 versions
in front of the SVP instead of one, from a five-team study done in one sitting.
Design target
$100M
a year for each point of network revenue on a $10B network: the prize for testing more versions of every route.
Estimated: arithmetic (1% × $10B), not a result
14 · What goes in and out
Your data in, a priced and checked route version out, with a person in between
In shortIt reads the data your teams already use and cites each number with its as-of date. What comes out is a route version you can defend, and a brief a leader can sign.
Demand and market
The DOT O&D survey and T-100 traffic
Your own bookings, fares and fare history
GDS booking data and settlement-based data
Every carrier's published schedules and fare shopping
Partner sales and onward flows
Operations and cost
The fleet register, lines of flying and the heavy-check calendar
The hub bank structure and published connecting times
Slot status and hub limits
Finance's cost model and fuel curve, with its version
Agents answer the whole studyDemand, feed, fleet, rivals, slots, partner and cost, on the same what-if
A person promotes and signs
Every number cited and dated
Outputs
Route versions with revenue, contribution and fleet fit, side by side
A brief with open disagreements and their sources
Drafted slot and codeshare requests for the teams that file them
Promoted changes sent to scheduling as SSIM
A signed season plan version
An audit trail and an evidence pack per decision
What changes for the team
Several versions in front of the leader, not one
The missing aircraft found before sale
A decision you can trace next season
15 · With what you already run
It works beside your planning suite, revenue management and schedule systems
In shortKeep your planning suite, revenue management, schedule and fleet systems. The new piece is the cross-team study around them, with a citation on every number and a person on every decision.
Your planning engine stays fast at the arithmetic. The agents run the cross-team study around it, cite every number, and route every decision to a person.
Network planning and schedule optimization suite
Agents read its scenario results and parameters where exposed, then explain, compare and challenge them. The optimizer keeps doing the math.
Revenue management system
Promoted plan changes flow out to it; fare and forecast context flows in to the study.
Schedule system and publication path
Agents read the schedule draft (banks, connecting times) and send promoted changes to scheduling as SSIM, the format it already reads.
O&D demand and traffic data
The DOT O&D survey, T-100, booking-channel and settlement-based data: each cited with its as-of date, and flagged when stale.
Published schedules and competitive data
Every carrier's schedule loads and fare shopping feed the competitive view, stated as a probability with its basis.
Fleet and maintenance planning
The fleet register, lines of flying and heavy-check calendar are read on every what-if; the fleet manager confirms.
Slot coordinator and FAA channels
Agents check slot status and hub limits and draft the request; the slot team files it.
Partner and alliance data
Partner bookings and codeshare terms value beyond feed; the DOT codeshare request is drafted for the alliance lead to approve.
Finance cost model
Routes are priced line by line on finance's own model and fuel curve, with its version; agents never change an assumption.
Three principles of agentic network planning
1
Run the study, not one calculation
Demand, feed, fleet, rivals, slots, partners and cost answer the same what-if together, and disagree in the open.
2
Agents propose; people decide
Agents compute, build versions and draft. Planners promote, the fleet manager confirms, the alliance lead approves, and the SVP signs, with a citation on every number.
3
Beside your engine, not instead of it
The tools you run stay in place. Agents read their outputs, add the cross-team checks, and send the promoted plan to scheduling.
Built on Botminds
One platform for agents, data and screens
Builder: describe the work, get a spec, build it, test it.
Runtime: agents, data, integrations, screens and models, in one place.
SecureIsolatedGovernedTestedScalableHighly available
Enterprise-ready
Roles and separation of duties: planners promote, the SVP signs, agents never file
Maker-checker on promotions, fleet findings and partner drafts
An unedited audit trail and exportable evidence packs
Integrations with planning, schedule, RM, fleet, slot, partner and finance systems
Single sign-on, service levels and support
Part C · Are you ready to automate?
Should you automate the new-route study? A buyer's view
You run the network, sign the season plan, or pay for the team that builds it. This part is written for you: what slow studies cost today, what automating them could return, whether your airline is ready, and what to ask any vendor, including us.
16 · What slow studies cost
What a slow or incomplete route study costs, before you buy anything
In shortNobody publishes the cost of a wrong route or a slow study, so start from your own numbers. Once you trust them, you can judge any solution against them, ours included.
$100M
a year for each point of network revenue on a $10 billion network. Small improvements in route decisions are worth a lot.
Illustrative arithmetic: 1% × $10B
2.0%
industry net margin forecast for 2026, down from 2025. There is little room for a route that loses money.1
IATA, June 7, 2026
0
published industry benchmarks for new-route study time, route-launch failure rates or the cost of a wrong route. Your own records are the only baseline.
See section 9
Find your own numbers
Cost
Where it hides
How to find your number in a week
Study time
Waiting for the next team's hand-off
Count the working days from idea to recommendation on your last few studies, and where each one waited
Team effort
Days spent across forecasting, schedule, fleet, slots, alliances and finance
Team-days per study, all teams together
Versions per decision
Only one version reaches the decision
How many versions your leader saw on the last five routes
Late fleet surprises
Leased aircraft, late starts, changes after sale
Events in the last two seasons, and what each one cost
Forecast misses
A rival matched and the load fell short
Forecast against actual on last season's new routes
Costs no calculator counts
Changes after saleUnder US rules, a passenger whose flight is cancelled or significantly changed, and who declines the alternative, is owed a prompt refund.1011
One versionThe leader chooses from what reached the desk, not from the options that ran out of time.
The unpriced downsideA forecast that assumes no rival response hides the case you most need to see.
Decisions re-arguedWhen nobody can trace which data month or fuel curve drove a decision, it is argued again next season.
Stale dataThe US ticket survey went monthly in July 2025: fresher data, and more of it to reconcile.2
17 · Your return
What automating the study could return for you
In shortPut in your own numbers. Every default is an illustrative assumption, not a benchmark or a promise: replace it with yours, and set a lever to 0 if it does not apply.
The value levers
Revenue and contribution
Better route decisions
More versions tested per decision, with rival response and network contribution priced in.
network revenue × revenue points gained
Reliability and asset use
No route promoted without an aircraft
Fleet gaps found before promotion avoid wet leases, late starts and changes after sale.
late fleet events avoided × cost per event
Planner capacity
Study throughput
Agents read, check and price; planners frame, question and decide.
studies × team-days per study × share removed × cost per day
Traceability
Faster, settled sign-off
An evidence pack per decision; fewer decisions re-argued next season.
decisions re-opened × hours per re-opening
Data freshness
Stale feeds flagged
Every number carries its as-of date; open studies re-run when new data lands.
studies re-run on release ÷ studies open
Competition
The downside priced in
A matched case on every forecast, with its basis, before the decision.
routes with a matched case ÷ routes decided
The calculator
Every default below is an illustrative assumption, not a benchmark or a promise. Replace each with your own number, and set a lever to 0 if it doesn't apply to you.
A · Decision qualityB · Late fleet surprisesC · Study effortPayback
Illustrative yearly value
A · Decision quality$0
B · Fleet surprises avoided$0
C · Study effort$0
Total$0
Team-days freed a year0
Payback (months)–
Formulas, per year: A = network revenue × revenue points gained. One point on a $10B network is $100M (arithmetic, not a result). B = events avoided × (wet lease or swap + revenue lost to a late start + rebooking and refund cost). C = studies × team-days per study × share removed × cost per day. Payback = annual cost ÷ (A + B + C) × 12.
No published benchmark exists for study time or for the cost of a wrong route: use your own measures (section 20, questions E1 and E4). Not counted: traceability, data freshness and faster sign-off.
Start here18 · Readiness check
Are you ready to automate? Nine questions, five minutes
In shortPick the answer closest to your airline today. Eight questions score 0 to 2; the last is for reflection. The meter adds them up as you go. Each question says why it matters and what a strong answer looks like.
1
Data
Do your planners have current O&D demand data (the DOT O&D survey, your own bookings and booking-channel data) in one place, each with its as-of date?
Why it mattersAgents cite what they can read. Dated, connected sources let every forecast say which data month it rests on, and flag a stale feed before anyone relies on it.
A strong answer"The monthly O&D survey, our bookings and our booking-channel data load to one store, and every table carries its as-of date."
2
Fleet
For any proposed route, can a planner see which aircraft it would take each week, including heavy checks and spares, before the route is promoted?
Why it mattersA fleet gap found after promotion becomes a wet lease, a late start or a change after tickets are sold. Found before, it is a choice between versions.
A strong answer"The heavy-check calendar and lines of flying are shared with network planning, and every study checks them."
3
Feed
Do you value a long-haul route on its network contribution (connecting feed, and what it takes from your own flights), rather than its local market alone?
Why it mattersAt a hub, connecting passengers can be a large share of a long-haul flight, and a new gateway can draw traffic from your existing ones. Local-market numbers alone mislead in both directions.
A strong answer"Every study shows route contribution and network contribution, with the feed by spoke city."
4
Competition
Does every route forecast state what happens if a rival matches, with the basis for that estimate?
Why it mattersA forecast that assumes no response hides the downside. Stating the response as a probability with its basis lets the decision-maker weigh it.
A strong answer"We keep rivals' launch history and show the matched case next to the base case."
5
Constraints
Are slot status (Level 2 or 3, 80/20 usage, historics), FAA hub limits and traffic rights checked during the study rather than after it?
Why it mattersA route that cannot get slots, rights or an approved time is not a route. Checking late wastes the study and can miss the slot calendar.
A strong answer"The slot team's rules and the FAA limits at our hubs are written into the study checklist."
6
Versions
Does your SVP Network Planning usually see more than one version of a route decision, with the differences explained?
Why it mattersThe value of planning comes from testing alternatives: another start date, frequency, aircraft or bank. One version means the comparison happened in someone's head, or not at all.
A strong answer"Two or three versions, side by side, with why each one wins or loses."
7
Decision rights
Is it written down who may promote a route to the season plan, who confirms the fleet, and who signs the plan version, with a reason recorded?
Why it mattersAgents should compute and draft; people should decide. Written decision rights tell you exactly where the approvals belong.
A strong answer"Planners promote with a reason code, the fleet manager confirms, and the SVP signs each version."
8
Traceability
Six months later, can you show which data, assumptions and disagreements drove a route decision?
Why it mattersWhen a route underperforms, the question is whether the forecast was wrong or the world changed. Without a record, the decision gets re-argued from scratch.
A strong answer"Each decision has an evidence pack: sources with dates, the versions, the disagreements and who signed."
9
Vision · for reflection
If your CCO asked tonight for three versions of a new long-haul route, each checked for demand, feed, fleet, rivals, slots and cost, what would have to be true to have them on the desk tomorrow?
Why it mattersYour answer lists the blockers. Each one is a candidate job for an agent: reading a source, checking a constraint, pricing a version or drafting a request.
A strong answerA short, specific list, for example "fleet calendar access, the partner feed, and one place to compare versions."
What your score means
0–6
Build the foundation
Start with data and decision rights: connect O&D, bookings and the fleet calendar with as-of dates, and write down who promotes, confirms and signs. Measure your current study time on a few routes.
7–11
Ready with groundwork
Connect the one or two missing feeds (often the heavy-check calendar or partner data), then run a measured pilot on a handful of real route ideas, with your people approving every output.
12–16
Ready to pilot
Run one season's new-route studies through agents beside your current process. Measure versions per decision, study days, fleet surprises and forecast against actual.
19 · Questions for any vendor
Sixteen questions to ask before you buy airline network planning software
In shortAsk every vendor the same questions, us included. The answers show how the work is really done: what gets checked before a route is chosen, what evidence you keep, and who is allowed to decide.
Ask
A good answer
A warning sign
What gets answered
Does it answer demand, feed, fleet, rivals, slots, partners and cost for the same what-if?
All of them together, on one version, in one sitting
One calculation; the rest still passes by hand-off
Does it check the heavy-check calendar before a version can be chosen?
A version short of an aircraft cannot be promoted
Fleet is checked after the decision
Does it price a rival's response?
A matched case on every forecast, as a probability with its basis
Rivals assumed to stand still
Does it count connection feed by bank, with our connecting times?
Feed recounted when the bank changes, using your published connecting times
A fixed connecting share for every bank
Evidence
Does every number carry a source and an as-of date?
Yes, and stale feeds are flagged before anyone relies on them
Numbers with no date
When two answers disagree, can we see both?
Both, cited, with what would settle it and when
One blended number
Can we export the evidence behind a decision?
An evidence pack per decision and an audit trail that cannot be edited
The evidence lives only in the vendor's screens
Data
Which demand data does it use, and how does it handle markets the US ticket survey misses?
The survey, T-100, your bookings, booking-channel and partner data, each cited
The survey alone
Does it use our finance cost model?
Your model and fuel curve, with their version; assumptions change only with a planner
Its own cost defaults
Control
Can the software promote, sign or file on its own?
No. It computes and drafts; people promote, confirm, approve and sign
"It auto-promotes the strong cases"
Who confirms fleet findings and approves slot and codeshare drafts?
Named roles, each decision recorded with a reason
Anyone with access
Your data
Where do our network, booking and partner data live, and who at the vendor can see them?
Isolated to your airline, with access logged
Shared environments, vague answers
Is our data used to train anything?
No, in writing
"Only to improve the service"
Fit and proof
Does it work beside our planning suite, revenue management and schedule system?
It reads their outputs and sends promoted changes to scheduling as SSIM
Replace your planning suite first
Will you run a pilot on our own route ideas, measured?
Yes: versions per decision, study days, fleet surprises, forecast against actual
Demo data only
What does the price depend on?
A clear unit, and what is included
A number that appears only after the contract
20 · How to start
How to start: five steps to a decision you can defend
In shortMeasure first, then pilot on your own route ideas, then decide with numbers. None of these steps commits you to a vendor.
Time your last few studies. Count the working days from idea to recommendation, and where each one waited.
List your late fleet surprises from the last two seasons: leased aircraft, late starts and changes after sale, with what each cost.
Gather the inputs for a handful of real route ideas: O&D data, bookings, the fleet calendar and the cost model, each with its as-of date.
Run one season's new-route studies beside your current process, with your people approving every output.
Decide with four numbers: versions per decision, study days, fleet surprises, and forecast against actual.
The full working-session questions: 39 questions in 9 groups
A
Data: O&D demand and traffic
A1
Do we use the DOT O&D survey in its new form (a 40% monthly sample since July 2025) or still the older quarterly series? Who reloads it, and how soon after release?
Why it mattersThe new series is denser and fresher 2. A study on last year's quarterly data misses recent shifts.
Strong answer · warning sign
Monthly loads with a named owner.
Nobody knows which series the forecasts use.
A2
How do we size markets where foreign carriers sell most of the tickets, which the DOT survey does not cover?
Why it mattersInternational markets can be badly undercounted from US-carrier data alone.
Strong answer · warning sign
A stated method and data source for each region.
"We scale up the DOT number."
A3
Do we use T-100 segment and market data for rivals' loads, and with what lag?
Why it mattersRivals' loads show whether a market can take more capacity.
Strong answer · warning sign
Loaded monthly, lag known.
Pulled by hand for each study.
A4
Which booking-channel data do we buy (GDS, settlement-based, direct), and what share of our own sales do they miss?
Why it mattersGDS data miss direct web, app and call-center sales, and that share keeps growing.
Strong answer · warning sign
Coverage measured against our own sales.
Coverage assumed to be complete.
B
Data: bookings, fares and schedules
B1
Is our own booking and one-stop history available to planners daily?
Why it mattersYour passengers already flying a market via another gateway are the best signal for a new nonstop, and the traffic it may take from you.
Strong answer · warning sign
Daily, by O&D.
A monthly extract on request.
B2
Do we have a fare-shopping feed for competing one-stop fares in the markets we study?
Why it mattersOne-stop fares set the price a new nonstop must beat.
Strong answer · warning sign
Yes, for every study market.
Screenshots in a deck.
B3
Do we receive every carrier's schedule loads daily, and see a rival's new route the day it is loaded?
Why it mattersA rival's load is the earliest evidence that settles a competitive disagreement.
Strong answer · warning sign
Daily, with alerts on our markets.
We learn from the trade press.
B4
Do we keep a history of rivals' responses to our launches: which they matched, and how many months later?
Why it mattersPattern is the honest basis for a response probability.
Strong answer · warning sign
A maintained launch-and-response log.
Memory and anecdote.
C
Data: fleet, maintenance and cost
C1
Is the fleet register (tails, configuration, seats) the system of record, and is it current?
Why it mattersSeat counts drive revenue; configuration drives range.
Strong answer · warning sign
One register, owned and current.
Several spreadsheets that disagree.
C2
Is the heavy-check calendar versioned and shared with network planning?
Why it mattersIt is the most common source of late fleet surprises.
Strong answer · warning sign
Shared, with version numbers.
Lives with maintenance and is emailed on request.
C3
Is the route cost model (fuel curve, block-hour rates, fees, ownership) owned by finance, versioned and dated?
Why it mattersWith fuel near a third of costs 1, the fuel assumption behind each decision must be traceable.
Strong answer · warning sign
Finance owns it; every study cites its version.
Each analyst keeps a copy.
D
Data: slots, rights and partners
D1
Do we track slot holdings and usage against 80% per series, including the summer 2027 minimum of 7 slots per summer series?
Why it mattersLosing historics at a coordinated airport is hard to undo 4.
Strong answer · warning sign
Usage tracked weekly per series.
Checked at the end of the season.
D2
Are the FAA limits at our US hubs (for example ORD and EWR through October 30, 2027) written into our planning rules?
Why it mattersA new departure needs an approved time inside the hub's limits 78.
Strong answer · warning sign
Limits encoded and checked in every study.
"The slot team will sort it out."
D3
Is there a list of traffic rights by country pair, and who maintains it?
Why it mattersRights decide which markets are open at all.
Strong answer · warning sign
A maintained list with an owner in government affairs.
Asked case by case.
D4
Do partner bookings and codeshare prorates reach planners, and how fresh are they?
Why it mattersBeyond feed can change a route's case; stale partner data should be flagged, not trusted.
Strong answer · warning sign
Daily exchange with freshness shown.
Quarterly files from the alliance team.
E
Process
E1
How many working days does a five-team new-route study take today, and where does it wait?
Why it mattersNo published benchmark exists. Your own baseline is the number a pilot will be measured against.
Strong answer · warning sign
Measured on recent studies, with the waits named.
"It depends," with no example.
E2
How many route ideas a season are framed, studied, promoted and launched?
Why it mattersThe funnel shows whether good ideas die in the queue.
Strong answer · warning sign
A count at each stage.
Ideas aren't logged until studied.
E3
At what stage is the fleet check done: before or after promotion?
Why it mattersAfter promotion, every gap is expensive.
Strong answer · warning sign
Before, every time.
After, when scheduling builds rotations.
E4
How often has a promoted route needed a wet lease, a delayed start or a change after going on sale?
Why it mattersEach event carries lease cost, lost revenue, and rebooking or refund work under DOT rules 10.
Strong answer · warning sign
Counted per season, with cost.
Not tracked.
E5
How do we re-run studies when new data lands, such as the monthly O&D survey or a rival's schedule load?
Why it mattersA decision made on old data should be re-checked before it is signed.
Strong answer · warning sign
Open studies re-run on each release.
Never; the study is a snapshot.
E6
How are rival responses modeled: one number, or scenarios with a stated basis?
Why it mattersScenarios make the downside visible before the decision.
Strong answer · warning sign
Base and matched cases, every time.
The response is ignored.
F
Governance and sign-off
F1
Who may promote a route, confirm the fleet, approve partner filings and sign a plan version?
Why it mattersThese become the roles and approvals in any system you run.
Strong answer · warning sign
A written matrix.
Whoever is in the meeting.
F2
Do promotions carry reason codes, and an attestation that sources and open disagreements were reviewed?
Why it mattersReasons make decisions comparable across seasons.
Strong answer · warning sign
A short list of reason codes, always used.
Free-text emails.
F3
Is there a service level for sign-off, with escalation when it is missed?
Why it mattersA late signature can miss a slot or FAA deadline.
Strong answer · warning sign
A stated service level with escalation.
Decisions wait for the next meeting.
F4
Is the decision record kept unedited, and can we export a decision's evidence for the CCO or the board?
Why it mattersAn evidence pack ends re-litigation and supports audit.
Strong answer · warning sign
Yes, in one export.
Rebuilt from inboxes.
F5
Which steps would we never let software do on its own, such as filing with a coordinator or DOT?
Why it mattersIt tells you where agents should draft and people should decide and file.
Strong answer · warning sign
Filings, promotions, sign-off and fleet commitments stay with people.
"All of it": then start with reading and checking.
G
Systems
G1
Which network planning or schedule optimization suite do we run, and can it export scenarios and results?
Why it mattersAgents work beside the engine: they read its results and run the cross-team study around them.
Strong answer · warning sign
Exports or an API exist.
Results live in screenshots.
G2
How does the schedule reach publication: SSIM out of the schedule system to distribution?
Why it mattersPromoted changes should flow to scheduling in the format it already reads.
Strong answer · warning sign
A known SSIM path and owner.
Re-keyed by hand.
G3
Which revenue management system consumes the promoted plan?
Why it mattersRevenue management sells what network planning decides.
Strong answer · warning sign
A defined hand-off.
RM learns from the schedule load.
G4
Which slot coordinator and FAA submission channels do we use?
Why it mattersDrafted requests should match each channel's format and calendar.
Strong answer · warning sign
A list per airport, with deadlines.
Known to one person.
G5
Where do fleet and maintenance planning live, and can they be read through an API or file feed?
Why it mattersThe fleet check depends on reading them every time.
Strong answer · warning sign
Feeds exist.
Screens, no export.
G6
What identity and access model must a new tool follow: single sign-on, roles, data residency?
Why it mattersKnowing the security review early avoids a late stop.
Strong answer · warning sign
A standard checklist and a named reviewer.
Unknown until procurement.
H
People
H1
How many network planners, route analysts and schedulers do we have, and how many studies does each carry?
Why it mattersCapacity, not ideas, is often the limit on how many versions get tested.
Strong answer · warning sign
Known, with a backlog count.
"Everyone is busy."
H2
Who is the internal owner, and who funds it: the SVP Network Planning, the CCO?
Why it mattersCross-team work stalls without one owner.
Strong answer · warning sign
A named owner and a budget line.
Nobody owns the study end to end.
H3
Which teams would see their work change: fleet, alliances, finance, slots, scheduling?
Why it mattersEach team's approvals and data must be designed in, not bolted on.
Strong answer · warning sign
Each team has a named contact.
"Network planning will just use it."
H4
Who would run a one-season pilot, and how would forecast accuracy be measured against actual results?
Why it mattersA measured pilot settles the decision.
Strong answer · warning sign
Forecast against actual, by route and month, plus study days.
"Let's see what it does."
I
Vision
I1
What if every route idea came back with three priced versions and the fleet, slot and rival checks done, before it reached the SVP?
Why it mattersTests appetite for more versions per decision.
Strong answer · warning sign
"We would test ideas we drop today."
Volume isn't the issue; find what is.
I2
What if every disagreement between teams arrived as a card with both sides cited and what would settle it?
Why it mattersLinks planning quality to how decisions are argued.
Strong answer · warning sign
"Our SVP would ask for that straight away."
Disagreements are settled before anyone senior sees them.
I3
Which route decision from last season would we most like to be able to replay, with its data and reasons?
Airline network planning is deciding which routes an airline flies, how often, on which aircraft and in which hub bank, season by season. The decision draws on O&D demand, connection feed, fleet and maintenance limits, competition, slots and traffic rights, partner flows and route economics.
How do airlines evaluate a new route?
They size the origin-and-destination market and the airline's likely share, add connecting feed through the hub, and subtract what the route takes from existing flights. Then they model a rival's response, check fleet, slots, rights and connecting times, and price the route line by line to reach route and network contribution.
What data does airline network planning use in the US?
The DOT origin-and-destination ticket survey (since July 2025, a 40% monthly sample of tickets issued by US carriers), T-100 segment traffic, the airline's own bookings and fares, booking-channel data, published schedules, the fleet and maintenance calendars, slot holdings and the finance cost model.
What is a hub bank, and why does it change a route's value?
A hub bank is a wave of arrivals followed by a wave of departures. A departure at the end of a bank can be fed by every spoke city whose arrival meets the minimum connecting time. Moving it one bank earlier or later changes how many cities feed it, and so its load and revenue.
How do airport slots affect a new route?
At slot-coordinated (Level 3) airports a route needs allocated slots. Under the IATA guidelines a series keeps historic precedence if it is flown at least 80% of the time, and from summer 2027 a summer series is at least seven slots. In the US, as of October 2026, the FAA controls slots at JFK, LaGuardia and Reagan National, facilitates schedules at Level 2 airports such as O'Hare, Newark, Los Angeles and San Francisco, and limits operations at O'Hare and Newark through October 30, 2027.
What happens to a route forecast if a competitor matches?
A good study shows both views side by side: the demand forecast without a response, and the forecast if a rival matches. The response should be stated as a probability with its basis, such as the rival's record of matching past launches, never as fact, along with the evidence that would settle it.
Does agentic network planning check fleet assignment and heavy checks before a route is promoted?
In the Botminds solution, yes. The Fleet agent checks every what-if against the fleet register, lines of flying and the heavy-check calendar, and proposes fixes from cheapest: move the start, use another aircraft, or wet-lease. The fleet planning manager confirms, and a version short of an aircraft cannot be promoted.
Do the agents file slot requests or codeshare applications?
No. Agents draft the slot request and the DOT codeshare authorization request with their cited facts. People review, approve and file them.
Does it replace our network planning or schedule optimization suite?
No. It works beside the tools you run. It reads their outputs and your data, runs the cross-team study around them, and sends the promoted plan to scheduling as SSIM.
Who stays in control of the network plan?
People. A network planner promotes a version with a reason code and an attestation, the fleet planning manager confirms the fleet, and the SVP Network Planning signs each plan version. Agents compute, build versions and draft, and never promote, sign, confirm or file. Every run and decision is kept on an audit trail with an exportable evidence pack.
What happens if a data feed is stale?
Every source shows its as-of date and status. A stale feed is flagged, and the answers that depend on it are marked until it is current. A retry re-checks the affected agents.
How do we get started with agentic network planning?
Take the five-minute readiness check in this guide, see the solution working on real screens, then try a walkthrough on one of your own route ideas. Go-live follows your requirements and data connections, and is tested before it reaches your planners.
22 · One-page summary
Airline network planning: one-page summary
Print this page
The readiness questions
Is our O&D demand data in one place, each source with its as-of date?
Can a planner see which aircraft a route would take, week by week, before promotion?
Do we value long-haul routes on network contribution, feed included?
Does every forecast show the case where a rival matches, with its basis?
Are slots, FAA hub limits and traffic rights checked during the study?
Does the SVP see more than one version, with the differences explained?
Are decision rights written down, with reasons recorded?
Can we show, six months later, what drove a route decision?
What would it take to put three checked versions of a route on the CCO's desk tomorrow?
Score 0–2 on 1–8: 0–6 build the foundation · 7–11 ready with groundwork · 12–16 ready to pilot.
12 terms
O&D
A passenger's true origin and destination.
DOT O&D survey
40% monthly ticket sample since July 2025 (DB1C).
T-100
DOT monthly segment and market traffic.
Load factor
Passengers ÷ seats.
Network contribution
Route contribution minus what it takes from own flights.
Hub bank
A wave of arrivals, then departures, for connections.
MCT
Minimum connecting time, by connection type.
Fleet assignment
Aircraft types matched to flights.
Heavy check
Maintenance that grounds an aircraft for weeks.
Level 3 airport
Slot-coordinated; every flight needs a slot.
80/20
Fly a series 80% of the time to keep it.
SSIM
IATA format for schedules and slot messages.
23 · Sources
Sources and dates
Every statistic in this guide links here. Regulatory statements are as of October 2026; rules change, so check the current text before relying on them.
IATA, "Middle East disruptions and high fuel prices halve airline industry profitability" (industry outlook), June 7, 2026. iata.org
US DOT, final rule "Updates to the Origin-Destination Survey of Airline Passengers," Federal Register, January 31, 2023 (40% monthly sample for travel from July 1, 2025). federalregister.gov
Bureau of Transportation Statistics: DB1C and the O&D survey data product; DB1B database profile (10% quarterly sample, historic series). BTS answers · TranStats
ACI, IATA and WWACG, Worldwide Airport Slot Guidelines (WASG), Edition 4, effective August 1, 2025. iata.org (PDF)
IATA, calendar of coordination activities (summer 2027). iata.org (PDF)
FAA, notice of submission deadline for schedule information, summer 2027, Federal Register, October 9, 2026. federalregister.gov
FAA, operating limitations at Chicago O'Hare: order of April 20, 2026, and extension of July 14, 2026. April order · July extension
FAA, operating limitations at Newark Liberty, extension, Federal Register, June 23, 2026. federalregister.gov
FAA, staffing-related relief concerning operations at Reagan National, JFK and LaGuardia, Federal Register, June 23, 2026. federalregister.gov
14 CFR Part 260, Refunds for airline fare and ancillary service fees (text current October 1, 2026). ecfr.gov
US DOT, final rule "Refunds and Other Consumer Protections," April 26, 2024; and enforcement notice of July 7, 2026. 2024 rule · 2026 notice
US–EU Air Transport Agreement (signed 2007, applied from March 30, 2008), European Parliament legislative summary. europarl.europa.eu
14 CFR Part 121, Appendix P, extended operations (ETOPS). law.cornell.edu
Klophaus, R. and Viehmann, J., on the survival of new air services, Journal of Air Transport Management, vol. 123, 2025. ideas.repec.org
Belobaba, P., Airline Management (16.75J), lecture notes on the airline planning process, MIT OpenCourseWare. ocw.mit.edu
On spill and recapture: Columbia Center for Pricing and Revenue Management, working paper 2009-2; MIT 1.206J Airline Schedule Planning, lecture 5. columbia.edu · ocw.mit.edu
Press report on Athens airport moving to Level 3 coordination for 2027, October 14, 2026 (confirm with the Hellenic Slot Coordination Authority). travelextra.ie
Consumer guide to airline booking windows (major US carriers open sales about 11 months out), 2026. thriftytraveler.com
US DOT, statements of authorization for codeshare services under 14 CFR Part 212, Federal Register, January 21, 1999. govinfo.gov
Airports Council International–North America, presentation on quality of service index (QSI) models, 2020. airportscouncil.org (PDF)
US DOT, U.S. International Air Passenger and Freight Statistics report (international T-100 data released after a six-month confidentiality period). transportation.gov
Next step
See it on one of your own route ideas.
Explore Botminds Airline Network Planning, or talk to us about a measured pilot on a handful of your own route ideas.
We use essential cookies to make this site work, and cookieless, anonymised analytics to understand how it's used. "Accept all" also turns on Google Analytics, which sets cookies. Privacy policy