The airline planning stack,
rebuilt as agents.
avint is a suite of AI agents for airline network, fleet and cost decisions. Each agent owns one domain, calls the others when it needs them, and shows every step of its work.
- aerocost.searchResolve the codes
- aerocost.coverageCheck what the answer will rest on before giving it
- aerocost.block_hour_costPrice the block hour on Lufthansa Group's own unit cost
- aerocost.cost_structureSplit it into the slices that move independently
- aerocost.fuel_scenarioRestate at today's jet fuel, because the published figure predates the step
Hallo, what is the cost of a 32N for LH?
Showing Lufthansa Group. Nine carriers are loaded, all from their own filings.
The Cost Model is the pricing authority for the whole suite. Ask it in plain language, get the figure, what it rests on, and how far to trust it. Every number traces back to a source, and an unknown is shown as unknown, never defaulted.
- Sabre
- MIDT
- Cirium
- Custom BI
- Consulting decks
- The spreadsheet
Airlines spend $40B a year on tools that don't talk to each other.
Distribution in one system, market data in another, schedules in a third, and the real analysis in a spreadsheet someone built in 2019. Every planning cycle starts by stitching these together by hand, and the stitching is where the weeks go.
The cost isn't the licences. It's the decisions that arrive late, the routes never evaluated because nobody had the time, and the signals that were sitting in one department while another one needed them.
One system that reasons across the whole airline.
Every avint agent is also a tool. The Cost Model prices a block hour for the Lease agent, the Lease agent constrains the Fleet agent, the Fleet agent feeds the Network agent. Each one you add makes the others sharper, because they share context rather than exports.
Every agent is a tool
Each agent exposes what it knows over a standard interface, so any other agent, or your own systems, can call it mid-reasoning. No exports, no copies of the numbers.
Every answer shows its trace
You see which tools ran, what came back, and how long it took. A figure you can't trace is a figure you can't defend in a board meeting.
Your cost base, not an average
Fuel at your region's price for your period. Non-fuel on your carrier's published unit costs. A narrowbody in Scandinavia does not cost what one in Texas costs, and the model knows it.
Where the operation loses time, from what actually flew.
ADS-B traces of every leg matched against the published schedule. Not a punctuality dashboard: it says which turnarounds are scheduled too tight, which routes are padded, how delay travels through the day, and what to change in the next schedule.
- schedule.loadLoad the published schedule and convert every leg to UTC
- adsb.matchMatch ADS-B traces to scheduled legs, reassemble split flights
- reliability.otpMeasure punctuality against the published times
- reliability.propagationTrace delay through every rotation to find where it is born
- schedule.evaluateTest block padding and turnaround times route by route
Where did we lose time in August, and what should we actually change in the winter schedule?
Return conditions, checked against reality.
Every lease clause tested against your live maintenance data, every shortfall priced by the Cost Model, and the whole portfolio ranked by money at risk per month you have left. Nothing rekeyed from a PDF.
- lease.load_portfolioLoad active leases and return condition clauses
- maintenance.syncSync maintenance status from your MRO system
- lease.check_conditionsTest every clause against current and forecast status
- cost_model.penaltyPrice each shortfall via the Cost Model agent
- lease.rankRank by exposure and time left to act
Which leased aircraft have return conditions at risk in the next 12 months, and what does it cost if we miss them?
Each one owns a domain. Three are live today.
Every agent below runs on the same trace-and-output pattern you saw above. New ones ship as the suite grows, and each one already knows how to call the others.
- Live
Cost Model
Localised block hour, delay cost by channel, Regulation 261 exposure and reactionary propagation, with provenance on every figure.
- Live
Fleet Reliability
ADS-B traces matched to the published schedule. Where time is lost, how delay propagates through rotations, and what to change.
- Live
Lease Management
Return conditions tested against live maintenance status, shortfalls priced, and the exposure ranked by how long you have to act.
- Coming
Route Planning
Evaluate new routes, simulate launch scenarios and see the ripple across the hub before committing an aircraft.
- Coming
Network Simulation
Model how adding or cutting a route cascades through the whole network.
- Coming
Fleet Management
Match types to routes and utilisation, with lease and cost constraints already in the loop.
- Coming
Pricing
Fare class and elasticity driven pricing that responds to market signals rather than last year's spreadsheet.
- Coming
Competitive Landscape
Competitor capacity moves, market share shifts and schedule changes, tracked as they happen.
- Coming
Loyalty & Revenue
Programme analytics, ancillary revenue and customer value, fed back into demand and pricing.
Airlines and the advisors airlines trust.
avint is built alongside operators and the strategy firms they already work with, so the agents answer the questions that actually come up in a planning meeting.
Bring us a decision you're working on.
Half an hour with someone who has done airline network planning for twelve years. Bring a real route, a real lease, or a real delay problem and we'll run it through the agents live.