AI and Optimization in Airline Pricing: Balancing Profit, Fairness, and Customer Trust

Image generated with Ai

In an exclusive interview with TTW, Jerry Yurchisin, Senior Data Science Strategist at Gurobi Optimization, shares insights from over a decade of experience in operations research, machine learning, statistics, and data visualisation to enhance decision-making.

Before joining Gurobi, Jerry served as Senior Consultant at OnLocation, Inc., where he customised linear programming models within the U.S. National Energy Modeling System (NEMS) to evaluate energy policies and emerging technologies. His earlier career included roles as Operations Research Analyst and Data Scientist at Booz Allen Hamilton, where he developed optimisation models and machine learning solutions for the U.S. Coast Guard, Department of Defense, and Department of Homeland Security.

Jerry holds both B.S. Ed. and M.S. degrees in Mathematics from Ohio University, as well as an M.S. in Operations Research and Statistics from the University of North Carolina at Chapel Hill, combining technical expertise with practical industry experience to tackle complex analytical challenges.

Airline pricing has evolved significantly over the decades, with mathematical optimization at the heart of the transformation. As far back as 1991, American Airlines won the Franz Edelman Award for its pioneering application of revenue, or yield, management to ticketing and operations. Today, optimization models remain powerful tools for revenue management, enabling airlines to align pricing strategies with business objectives—whether that means maximising profits, filling seats, or achieving a balance. These models can integrate numerous constraints, from capacity limits to competitive dynamics, to identify the best trade-offs for the airline’s strategic goals.

The emergence of AI-driven pricing introduces both opportunities and challenges, especially in consumer-facing industries like aviation. To ensure transparency and fairness, companies should clearly communicate which factors influence pricing decisions and which do not. Customers should have channels to ask questions or raise concerns, while key fairness-related business rules—such as caps on fare increases or anti-discrimination measures—must be embedded as non-negotiable “hard constraints” in optimization models.

One of the biggest challenges in implementing AI-based pricing is over-reliance on historical data, which may not reflect future market conditions in a volatile sector like air travel. To avoid this, airlines should supplement past data with real-time external factors such as economic indicators and customer sentiment. Maintaining transparency about pricing priorities is crucial for preserving trust, as opaque or profit-only models can damage brand loyalty.

Predictive modelling plays a central role by forecasting factors such as demand, no-show rates, and competitor behaviour, feeding these insights into optimization systems to inform pricing strategies. However, risks include bias embedded in historical data, inaccurate parameters, overfitting, and dangerous feedback loops—where one airline’s price change triggers competitor responses that spiral prices upward. The benefits, though, are substantial: greater efficiency, less waste, better supply-demand alignment, and potentially lower fares.

When deploying advanced pricing systems, companies should first define clear objectives, then implement guardrails like maximum price caps or fairness rules. Optimization can ensure these safeguards are enforced automatically, preventing violations.

Looking ahead, AI’s growing influence is expected to reshape the airline industry’s pricing dynamics and competitive landscape. An “arms race” for faster, smarter models is likely, driving more personalised pricing despite potential regulatory pushback. Price dispersion may widen, creating both higher peaks for some passengers and lower fares for others. This environment could boost the relevance of third-party comparison sites as customers seek transparency. Meanwhile, airlines will face increased pressure to offer meaningful perks to retain loyalty, particularly if travellers feel they are consistently paying more. The future of airline pricing will hinge on the industry’s ability to balance technological innovation with fairness, clarity, and customer trust.

How has airline pricing evolved over the years, and how does mathematical optimization currently balance customer demand with business constraints?

Airline pricing has a long history of using optimization. In fact, back in 1991, American Airlines won the Franz Edelman Award for applying revenue (or yield) management to ticketing and operations. Today, mathematical optimization remains one of the most flexible and powerful tools in airline revenue management. It allows companies to tailor models to their specific business goals, whether that’s maximizing profit, filling seats, or striking a balance between the two. The model can incorporate dozens of constraints, such as capacity limits, competitive pressures, and operational requirements, and then find the best possible trade-offs that align with the airline’s strategy.

With the rise of AI-driven pricing, what steps can businesses take to ensure transparency and fairness, particularly in consumer-facing industries like airlines?

Transparency starts with clarity. Companies should openly communicate which factors matter in pricing decisions and which don’t. For example, if competitor pricing influences your model, say so. There should also be clear avenues for customers to ask questions or voice concerns.

From a modeling perspective, certain business rules, like caps on fare increases or prohibiting discrimination, should be “hard constraints” in the optimization model. That means they can never be violated, ensuring fairness and consistency.

What are the main challenges airlines face when implementing AI-based pricing models, and how can they overcome these obstacles to maintain customer trust?

One of the biggest risks is relying too heavily on historical data. The past isn’t always predictive of the future, especially in a volatile industry like air travel. Airlines should incorporate external factors such as customer sentiment or macroeconomic trends to avoid tunnel vision. Ultimately, transparency is key, showing what the airline prioritizes in its pricing models can help maintain trust. If customers feel the process is opaque or purely profit-driven, it risks eroding brand loyalty.

How does predictive modeling play a role in pricing, and what are the potential risks and benefits associated with its use in an industry like aviation?

Predictive modeling allows airlines to estimate key factors like flight demand, no-show rates, and competitor reactions. These forecasts then feed into the optimization model to determine the most effective pricing strategy. However, there are risks: models can have inherit bias from historical data, parameters can be incorrect, and overfitting can make them overly sensitive to small changes in input.

Feedback loops are another danger, if one airline raises prices and competitors follow, it can spiral upward quickly. The benefits, though, are significant: improved efficiency, reduced waste, and potentially lower fares, plus the ability to deliver better customer experiences by aligning supply and demand more effectively.

What key factors should companies consider when deploying advanced pricing systems, especially when the system impacts consumer decisions and overall market fairness?

First, set clear objectives; what’s the main goal of the pricing model? Second, establish guardrails, such as maximum price thresholds or fairness rules. Mathematical optimization is particularly well-suited to ensuring these constraints are baked into the decision-making process and can’t be violated.

As AI continues to influence the airline industry, what long-term implications do you foresee for both businesses and consumers in terms of pricing dynamics and market competition?

We’re likely to see an “arms race” where airlines compete on who can deploy better, faster models. Pricing will likely become more personalized, despite the pushback this approach sometimes generates, unless regulations limit it. Prices may also become more widely dispersed, with higher highs for certain airlines and lower lows for others. This could increase the value of third-party comparison sites, as customers seek transparency. Brand loyalty will also be tested, if customers perceive they’re consistently paying more, airlines will need to offer meaningful perks or benefits to keep them coming back.