We word every claim at the evidence behind it.

No. Every KBM Air product runs on synthetic fixtures or public data feeds. No product is deployed in a customer’s live operation, which is why nothing on this site is labelled OPERATIONAL. When that changes, the evidence page changes first and it will name the tier, the date and the method.

None that we can publish, and we do not publish ones we do not have. There are no customer logos, testimonials or outcome figures anywhere on this site for that reason.

From seeded synthetic data, except where a product ingests a public feed — aviation weather and published navigation data are real. A cost, delay or misconnection figure inside a demonstration describes the demonstration. It is not a result from an airline and is never presented as one.

Because none has been verified against its primary source from this workspace. The content model requires a resolvable source URL, the publication year, the caveat the figure carries and the date the wording was last checked before a statistic can be rendered at all. An unsourced number on an aviation site is the first thing a reviewer checks, so we would rather publish none than publish one we half-remember.

Because leaving it off would misrepresent the portfolio. It is labelled CONCEPT, its status line says nothing has been built or demonstrated, and its claim register contains no DEMONSTRATED entry at all. A vendor who only shows you their strongest product is telling you something about the others.

No. It is not FAA Flight Service, holds no approval or acceptance from any authority, issues no go/no-go verdict, and is not a substitute for an official briefing. It assembles public weather sources into one route-shaped view and labels which sections are live. The pilot in command retains full responsibility for every flight safety decision.

Notices to airmen are placeholder strings and special use airspace is not queried at all — both are badged on the surface of the briefing, not disclosed in a footnote. Surface weather, winds and temperatures aloft, route-filtered adverse conditions and temporary flight restrictions come from live public sources. Geolocation is captured but no output depends on it yet.

Not until a person has verified the performance table for that airframe against the aircraft flight manual. The module marks every unverified table unusable and refuses to extrapolate beyond the published envelope, reporting "outside the published data" rather than producing a plausible number. A wrong takeoff distance is not a display bug.

No. Every safety-relevant function is decision support. Nothing releases an aircraft, files or amends a flight plan, publishes a weather briefing, or rebooks a passenger without a qualified human approving it — and the approval is recorded as a human act, with the person, the time and the reasoning.

No. Safety-relevant text and scoring come from explicit rules. In AeroVox the entire text path is deterministic — the model is only the voice, and removing it leaves a complete briefing in text. In the operations products the analyst answers from the records, and with no model credential configured at all it still answers.

No. It is not connected to any transmitter, is not an ATIS, AWOS or ASOS replacement, and nothing it produces goes out on an aviation frequency. It renders an authoritative observation into a reviewable briefing — text and audio together — for the briefing channels an operator already runs, with a qualified person reviewing before distribution.

Each product has a defined degraded behaviour and they are published on the trust page. In short: no model credential means the deterministic path answers anyway; an unreachable weather service produces a clearly labelled fallback or no new briefing rather than stale data presented as current; a missing record produces a named gap rather than an interpolated answer.

No, and we do not claim to be. No aviation authority has certified, approved or accepted any of this software. SOC 2 and FedRAMP alignment are costed roadmap items driven by procurement need; neither programme has started. The trust page lists everything we do not hold, which is the more useful list.

Nothing, because none of it is built — no profile has been captured, no position read and no camera opened, as there is no application. A real build would need explicit granular consent, a stated retention position and a clear answer on whether an employer could ever see a crew member’s location; our starting position is that they could not. That consent design is a precondition of any pilot, not a later refinement, and it does not exist yet.

No. Adapters are written against published schemas and run as clearly labelled simulations. Supplying credentials switches the same code paths to live services, but no such relationship exists today and none is implied by the schemas we build to.

Yes. The demonstration surfaces are self-contained with relative paths and no mandatory external service, so they run on a dedicated host or inside a controlled environment. Secrets stay in the operator’s environment configuration, and the deterministic answer paths mean no outbound model call is required for the product to work.

No. There is no customer data today because there is no customer deployment, and nothing in the current build trains, fine-tunes or retains anything for model improvement. Any future arrangement would be a written term, not a default.

Rebound, because reaccommodation input is a passenger manifest — names, parties, accessibility and medical needs, loyalty status. Today every manifest is synthetic and no real passenger data has ever been in it. AeroVox holds no personal data at all. Each product page states its own boundary.

A scoped engagement against your operation’s reality, with the success criteria written down first and the evidence tier of every claim agreed before it starts. We would rather define what would count as a failure at the beginning than argue about it at the end.

AeroVox, because it settles its own argument: the same live observation rendered by legacy automation and by AeroVox, back to back, with both transcripts on screen. You do not need a briefing to judge it.

No — regional carriers and Part 135 operators are the buyers this portfolio is shaped around, not a scaled-down version of a major-carrier product. The more useful question to press us on is deployment footprint and price, because those are what make enterprise tooling unusable at your size.