AI Policy

The short version

Bravo is a directory and review platform. The ratings, reviews, safety tiers, and registry facts you see are records, not generated text. No review on Bravo is written, scored, or ranked by AI. No decision about a person or a company is made by an algorithm alone. Where software assembles text from data, this page says so.

What is never AI

• Reviews. Every review is written by a person: a Google reviewer, or a Bravo member who confirmed a flight. We do not write, rewrite, summarize into new reviews, or invent reviews.
• Ratings and rankings. Star ratings are averages of real reviews. The default “Top rated” order is a volume-weighted average (a Bayesian average that stops a 5.0 on two reviews outranking a 4.8 on four hundred), with ties broken by review count, then fleet size, then name. Sponsored placement is labeled. There is no machine-learned score.
• Safety ratings. ARGUS and WYVERN tiers come from those auditors’ official registries and are shown as published.
• Registry facts. Tail numbers, aircraft types, airport data, and runway figures come from the FAA registry and public aviation datasets.

Where software assembles text from data

• Review themes. Cards such as “Crew & professionalism” or “On-time & reliable” count how many stored reviews mention keywords for that theme and pull one real sentence from a real review as the example. This is keyword matching over text people wrote. Nothing in a theme card is generated, and every count and quote traces to a stored review.
• Airport overviews and FAQs. The overview paragraph and the question-and-answer block on an airport, route, or state page are built from fixed sentence templates filled with that page’s data (location, elevation, runways, based operators, nearby airports). No language model is involved. The wording is stable from build to build.
• Trip concierge and cost estimator. When you type a trip in plain language, a rule-based parser picks out the airports, date, and party size and matches an aircraft class from published performance data. It is not a chatbot and does not send your text to an AI service.

Editorial content

Guides and blog articles are written and edited by the Bravo team. We may use AI writing tools to draft or tidy that editorial copy. A person reviews every article before it is published, checks the facts against our data, and is responsible for it. Editorial content never becomes a review or a rating.

No automated decisions about people

Listing claims, company submissions, review disputes, and account actions are decided by a person on the Bravo team. A few fixed rules run when a review is submitted and can decline it on the spot: you manage the listing you are reviewing, you reviewed the same company in the last 90 days, the review contains web links, or you are posting reviews very quickly. You see the reason immediately and can fix it. These are fixed rules, not models; they never remove a published review or close an account. Nothing on Bravo profiles you or predicts your behavior.

Third parties

Google Analytics, Google AdSense, and Google reCAPTCHA run on parts of the site and use their own automated systems for measurement, ad selection, and bot detection, under your cookie choices. They do not receive review content or quote details. See the Privacy Policy for what each one collects.

AI crawlers and our data

Public register pages may be read by search engines and AI assistants so that they can cite Bravo for operator, broker, and airport facts. Private areas (accounts, dashboards, quote threads, claims) are blocked in robots.txt and require sign-in. Reviews and ratings on Bravo are the work of their authors and may not be scraped in bulk or republished without permission.

Changes

If we start using a language model anywhere in the product, this page will say where, what data it sees, and how a person stays in the loop, before that feature goes live. Questions: contact us.