Krila AI

Krila AI — A joined airline planning layer plus its AI decision outputs: the platform connects to an airline's PSS and to OAG and ATPCO feeds, then joins schedules, bookings, competitive capacity and demand signals at origin-destination level in one always-current store — and, as byproduct, retains the Krila Score route rankings with their explicit factor decompositions (demand, connectivity, competitive gap, operational fit), the what-if scenario runs with their contribution/load-factor/aircraft-availability sensitivities, and the drafted competitor-response options with attached confidence …

Krila AI offers (Alternative, Price, Reference, Fundamental) — Small today and largely synthetic in the public record — the visible evidence is worked examples rather than a published corpus (a 742-sector-style twin is NOT claimed here). The underlying join is at O&D level across an airline's network, so per-customer volume scales with that carrier's O&D count and booking horizon; total platform volume is currently bounded by a very small deployment footprint..

Training airline network-planning agents (state = network + fleet + competitor filings, action = schedule change, reward = modelled annual contribution); route-launch candidate ranking models; capacity-deployment and aircraft-assignment RL environments with codeshare and hub-bank dynamics; fare-class-mix and RASM-gap attribution models; competitor-response prediction and simulation (given a rival filing, what response maximises margin); schedule-change event-response datasets with minute-level reaction timestamps; aviation natural-language-to-analytics corpora from the query layer ('show me capacity vs bookings on every transcon route'); agent-evaluation sets for explainability, since every recommendation carries visible assumptions.

Coverage spans US, UK, Europe; Passenger Airlines, Application Software; alternative, price, reference, fundamental; equities.