The intelligence
Intelligence with a record of why.
Deterministic reasoning and versioned legal knowledge, made useful for everyday decisions.
Reasoning you can inspect
Preflight’s assessment engine uses deterministic algorithms and expert-authored legal knowledge. The same facts, methodology and legal date produce the same analysis. Findings retain their supporting sources, dependencies and unresolved judgments.
Reading a description without a model
The activity description is read with a versioned vocabulary and clause grammar that understands negation, hedging and contradiction. Each proposed observation cites the exact words it came from, and it is only recorded once a person confirms it.
Exact about what is still open
For every unanswered fact the rating depends on, Preflight considers each possible answer and reports the lowest and highest rating any of them could produce. That is how it knows which questions matter, and why it never shows an open assessment as Low.
Facts are not permission
An operator can describe a purpose, an access arrangement or the steps taken. Those observations do not automatically establish lawful handling, valid consent or reasonable steps. The product preserves that distinction.
No model calls in the assessment engine
Preflight evaluation does not use runtime language models, embeddings or external generative inference services. Every result can be traced to the rules and facts that produced it.
Professional judgment remains essential
The method and every rule in the guided release are reviewed and approved by a qualified Australian lawyer, and each approval is bound to the exact wording reviewed: change the wording and the approval no longer applies. Sources are re-checked on a 90-day cycle, and a release goes public only when every rule in it is approved. Preflight is decision support, not legal advice.
Free to assess · Pay only for the PIA document · See pricing
Questions? Contact us