Cloak

Method

How they price you — and how Cloak answers

In August 2026 Delta’s CEO said AI would lift profits by about 50% by setting a different ticket price for every passenger in real time. The system is Fetcherr. Delta told Congress it was not using personal data for individualized pricing. Its investor deck described offers for “you, the individual.” Airlines are exempt from FTC pricing oversight.

What the model actually scores

  • A unique fare per shopping request — not a published bucket.
  • Repeat searches and logged-in sessions as “ready to buy.”
  • Device, cookies, time of day, and originating city as willingness-to-pay proxies.
  • Direct airline site vs a meta/OTA hop.
  • The same pattern already measured on Uber/Lyft: two people, same ride, $76 vs $24.

What Cloak sends instead

  • Eight shadow identities in parallel: fresh ghost, Eastern EU VPN, Mexico City, late-night mobile, six-month planner, elite, Bay Area, and “you.”
  • Nearby airports treated as different markets (JFK/EWR, LHR/LGW).
  • Hotels and cars run through the same scorer — member rate vs opaque, prepaid vs counter.
  • Vacation packages unbundled, ghosted per leg, then stacked.

What this desk is — and is not

Cloak does not scrape airline sites, break booking systems, or invent identities to defraud a carrier. It runs a Fetcherr-class willingness-to-pay model against a live inventory simulation, from eight shoppers at once, and books the lowest quote through the identity that received it. Point the same shadow desk at a GDS and the shopping logic does not change. Confirmations on this preview are simulated; your profile, two-factor, preferences, and screeners are real and stored per account.

Run a shadow shop