Measured against control, stated with scope.
Every result here carries its scope, period and comparison.
of AI responses about travel and loyalty brands are misaligned
Based on Bonafide data since 2024
Measuredrevenue advantage over untreated control, 99% statistical confidence
Global hotel group · 20 brands · six months
MeasuredAI accuracy in the first five months
International airline
Measuredshopping uplift against untreated control
International airline · 18 tracked routes · three-month baseline vs follow-up
Measuredgrowth in GenAI web sessions on test city pairs
Low-cost carrier · five months
Benchmarkof answers about airline products, policies and programs wrong before alignment
Bonafide airline benchmark
Bonafide measures context quality, recommendation behavior, qualified referrals and commercial trends. Wherever possible, results compare aligned properties or routes with an untreated control group. Booking-level attribution is used only when the parties have an agreed closed-loop tracking method.
Accuracy is reported as correct answers divided by all answers scored: correct, incorrect and unanswered. Unanswered questions count against the score.
Citation data is a useful diagnostic, but we don't use it as a headline result. What matters is whether AI understood you, recommended you and routed the customer correctly, and what that did commercially.