AI Helps Delta Shift 20% of Ticket Pricing to Real-Time Automation by 2025
By Rahul Neel Mani (@rneelmani)
One of the top three US airlines, Delta Air Lines, is taking a significant step with the help of artificial intelligence to transform its ticket pricing, a long-standing pain area of the airline industry. By the end of 2025, the airline plans to use AI to determine prices for up to 20 per cent of its fares, up from the current 3 per cent managed by machine learning.
The technology, developed by the Israeli startup Fetcherr, aims to find the best fare for each customer in real time. This bold move combines the potential of advanced optimization with the risks of eroding customer trust and government oversight. US senators are raising concerns over what some describe as “surveillance pricing,” but Delta insists it’s simply advancing traditional dynamic pricing with AI. If successful, the move could not only increase profits but also mark a new era of AI-driven personalization in travel.
“Super Analyst” in the Cloud
For Delta leadership, this transformation is a multi-phase journey. “We’re letting the machine go ahead and set prices in a very controlled way. We’re excited about partnering with Fetcherr. This is a full reengineering of how we price and how we will be pricing in the future,” Delta President Glen Hauenstein said during the Q2 2025 earnings call.
Initially launched as a pilot covering just 1 per cent of fares in late 2024, the AI system is expanding rapidly. Delta CEO Ed Bastian, while addressing investors, described the early results as “encouraging” and said that the airline plans to monitor the behavior of the AI model during the ramp-up phase. “We’re in a heavy testing phase. The more data we ingest, the more it learns,” he said.
What sets Delta’s approach apart is its goal to offer personalized pricing on a large scale. “We will have a price available on that aircraft at that time that’s accessible to the customer,” Hauenstein said. This means that if two travelers are searching for the same flight, they may see very different prices, based on variables ranging from historical behavior to their loyalty status.
This AI-led pricing initiative aligns with Delta’s broader business strategy. The airline reported a record US$15.5 billion in revenue for the June 2025 quarter and maintained an operating margin of 13 per cent. This certainly underscores its confidence in long-term digital investments.
Inside Fetcherr’s Generative Pricing Engine
What’s enabling Delta’s pricing? It’s an Israeli startup Fetcherr’s proprietary generative pricing engine, or GPE, being deployed. Fetcherr CEO Roy Cohen describes the technology as “the first AI hedge fund for airline revenue management.” Built on the company’s generative AI large language model, it processes a mix of internal airline data and external market signals ranging from weather and holidays to competitive pricing and macroeconomic indicators.
According to Fetcherr, the algorithm can simulate millions of pricing scenarios in milliseconds, forecast demand surges and also adjust fares accordingly across an entire flight network. All this, without any human intervention.
Traditional revenue management systems rely heavily on historical data and static fares. In contrast, Fetcherr’s approach is to utilize machine learning for real-time predictive decisions. This means pricing isn’t just reactive, it’s anticipatory. It’s like having a digital analyst, who never sleeps and understands the market. Fetcherr advises airlines on digital transformation.
Importantly, Fetcherr’s system can be integrated into existing airline infrastructure without requiring any overhaul. The onboarding process is quick and suitable for testing, and its pricing decisions include explainability features, a must for adhering to AI transparency.
The Privacy Pushback
The technology may show its benefits, but not everyone is convinced. US Senators Ruben Gallego, Mark Warner and Richard Blumenthal have written to Delta, expressing concern about how the AI will set fares. “This isn’t fair pricing or competitive pricing. It’s predatory pricing,” Senator Gallego said.
The primary concern raised by lawmakers is whether the AI-powered pricing system uses personal data to identify a customer’s pain point and manipulate fares accordingly. Critics argue that such capabilities could open doors to exploitative pricing in the future. But Delta has strongly denied this allegation. “There is no fare product Delta has ever used, is testing, or plans to use that targets customers with individualized offers based on personal information,” the company said.
Complications persist. Critics point to past cases where tech-enabled surge pricing led to customer and data exploitation, such as Orbitz charging Mac users more for hotel rooms or Target changing prices based on location data. The Federal Trade Commission has referred to this practice as surveillance pricing.
In Europe, where General Data Protection Regulation enforces stricter data use laws, such models could encounter significant challenges. Airlines need to demonstrate a legitimate interest in personalizing pricing, something that might not align with AI’s black-box logic.
What It Means for the Future
Despite the controversy, one thing is clear: Delta’s move will attract attention across industries. Other US airlines, such as United and American, have adopted AI for revenue forecasting. Fetcherr’s other airline clients include Virgin Atlantic, WestJet, Azul and Viva. The airline sector may well become a testing ground for AI-driven revenue optimization and other industries may follow soon.
The business case is strong. Fetcherr claims its AI engine has resulted in a 10 per cent revenue growth for its customers over the past three years. McKinsey estimates that AI could boost global corporate profits by US$4.4 trillion annually. In dynamic industry sectors, such as aviation, even small pricing improvements can lead to billions of dollars’ worth of savings.
For digital leaders, this isn’t just another AI experiment to watch from a distance. The architecture of pricing, trust and personalization is undergoing massive transformation. The key point isn’t about governance checklists or ethical frameworks. Instead, decision-making is gradually being handed over to AI systems that follow rules based on logic. In this approach, AI isn’t merely supporting the business; it’s a core part of the strategy.
“We’re all in on this,” Hauenstein said. “But we want to be smart about it, because it could also be hazardous if not done correctly.”
