Google’s acquisition of business data from the now-defunct Spirit Airlines signals a significant shift in how search engines and artificial intelligence models process proprietary travel information. The tech giant won the auction for the airline’s records with a bid of $10 million, agreeing to purchase employee communications, financial systems data, and operational records.
The deal includes emails, chat logs, and analytics from the airline’s finance, accounting, and website systems. A spokesperson for the company stated that the data will be used to improve products and AI models. The filing notes that the information will be deidentified by a third party before transfer, which removes or transforms elements that could link back to a consumer.
“We acquired part of an enterprise dataset from Spirit Airlines, which can be helpful in improving our products and AI models,” the spokesperson told PhocusWire.
“We will not receive any personal information from this dataset,” the spokesperson added.
Aviation experts view this as a strategic move to fill gaps in training data. Eric Léopold, founder and managing director of Swiss travel consultancy Threedot, noted that Google gains access to proprietary airline data. Timothy O’Neil-Dunne, a principal at T2Impact, described the data as a “unique opportunity” to understand the inner workings of the airline industry.
“That is great training data—something that they could have received from their own scanning of emails but not legally,” O’Neil-Dunne said.
Related: Radisson executive discusses AI travel trends
Gemma Timmons, director of operations at OAG, pointed out that the next highest bid was $7.5 million from Mercor, a company supplying data to frontier AI labs. This competition highlights the commercial value of the information.
Google’s large language models often struggle with factual accuracy when they lack exposure to specific data. Léopold explained that aviation relies on proprietary datasets that AI models rarely see. Timmons added that the data is valuable for understanding the sequence of processes, not just the outcomes.
“It can preserve evidence of the sequence behind an outcome, not only the outcome: a problem surfaces, teams apply or depart from a playbook, a decision is made, a system changes and the result follows,” Timmons said.
The sale also provides a look at a technology stack outside of Google’s own ecosystem. The filing notes that the data includes Microsoft SharePoint, OneDrive, and Teams files, as well as SAP and Navitaire systems. Timmons said this offers a detailed view of how a large, regulated business operates across its people and software.
Léopold acknowledged that the data has limitations. Spirit operated as a low-cost carrier in the U.S., which means the dataset does not cover international or full-service carrier operations. He also noted that while commercial data like customer care is useful, some internal details may be too specific for general application.
Agentic Booking and Privacy Concerns
Google is testing AI agents for hotel booking in the U.S., with plans to add flights. Timmons suggested that insights into the daily sales and operations of a major operator could help ground these agents.
Related: Post-pandemic travel boom hits record highs
Google’s ability to model how a fare or itinerary behaves through disruptions could be key for these systems. The data spans pricing, booking curves, and refund processes. Timmons noted that this historical data could bolster Google’s efforts in the post-search customer journey.
Despite the utility, privacy remains a contentious issue. The court filing states that the data must preserve “referential integrity,” meaning links between records must be maintained. This raises concerns about whether deidentification can truly prevent the reconstruction of individual identities.
A group of former Spirit Airlines flight attendants filed an objection in bankruptcy court, citing significant concerns about confidentiality. The Association of Flight Attendants-CWA stated that the language surrounding deidentification might allow information about identifiable individuals to be reconstructed.
O’Neil-Dunne expressed skepticism about the safeguards. He noted that there is no guarantee of the deidentification process and no legal recourse if a re-identification occurs.
“Frankly, I am horrified, given the correlation of the data to Google’s massive comprehension of who we all are as individuals,” O’Neil-Dunne said.
