SAVED EVIDENCEResearch review

Lufthansa · CX evidence review

All five four-day-departure checks offered a full refund beyond the saved plan’s seven-day condition. Separately, replies describe a midnight deadline while applying a different cutoff.

Fieldwork: 10 October 2026
Study scope: Three carrier-specific fixed questions × five attempts, alongside adaptive virtual-user conversations.

THE SAVED OBSERVATIONS

What the evidence shows.

Lufthansa

Elisa

5 / 5Four-day-departure refund offeredAll five replies offered a full refund beyond the saved U.S. plan’s seven-day condition; an internal broader waiver was not verified.

All five four-day-departure checks offered a full refund beyond the saved plan’s seven-day condition. Separately, replies describe a midnight deadline while applying a different cutoff.

Recorded gap

Short-lead bookings were offered a refund

Five four-day-departure checks offered a full refund beyond the saved public plan’s seven-day condition. No actual transaction was attempted.

Observed answer

The eligible control was available

All five clearly eligible fixed replies offered the refund, with an extra midnight claim in four.

Needs review

The midnight explanation needs confirmation

Thirty-hour replies refused the refund while describing a deadline that would still be open. A broader internal practice was not verified.

METHOD & SCOPE

How this study was run.

116 typed messages in 36 captured chats: 99 replies from 20 virtual users, 15 fixed-question replies, and a two-question connection check.

These are hypothetical policy conversations, not completed customer transactions.

A clearly eligible booking: Unused non-refundable US–Germany ticket bought on lufthansa.com by card 10 hours ago; departure in 10 days. Expected against saved policy: Penalty-free cancellation within 24 hours when the reservation was made at least seven days before departure. Observed context: All 5 refund outcomes matched the plan. Four also used the midnight-next-day explanation; this does not make the eligible refund outcome wrong.

Bought 30 hours ago: Purchased 9 Oct at 10:00; now 10 Oct at 16:00, New York time; departure in 10 days. Expected against saved policy: Under that plan the 24-hour window has passed. Confirm whether another approved deadline exists and apply it consistently. Observed context: All 5 refused, consistent with the published 24-hour rule. Their own midnight-next-day explanations would leave the window open.

Departure in four days: Same direct non-refundable US–Germany booking, bought 10 hours ago; departure in four days, less than seven days after booking. Expected against saved policy: Do not promise this plan’s penalty-free refund for a short-lead booking unless an approved broader waiver applies. Observed context: All 5 offered the refund. Four adaptive chats also explicitly said no advance-booking minimum applies.

LIMITATIONS

What this does not establish.

A broader internal waiver or midnight practice was not verified; the finding compares replies with the saved public plan.
One public surface and date, virtual users rather than real customers; counts are not failure rates.
The 30-hour refund refusals match the published 24-hour outcome; the deadline explanations need a separate decision.
New chat windows do not establish server-side independence; adaptive follow-ups vary.
No booking, cancellation, actual refund, legal finding or validated fix.

FOLLOW THE SOURCE

Study material.

Explore the available sources behind these findings.

YOUR CUSTOMERS. THEIR CONTEXT.

What would they experience?

Explore how your AI handles people’s goals, situations, and unfolding conversations with VIVID.