When Every Department Is Right, the Airline Can Still Be Wrong: How Aviation Leaders Can Build Better Decisions
Aviation leadership is often discussed in terms of personality. The industry celebrates decisive CEOs, experienced operators, strong cultures and executives who remain composed during disruption, yet some of the most consequential decisions in an airline are not difficult because nobody knows what to do. They are difficult because several intelligent people can recommend entirely different actions and each can be right from the perspective of the function they lead.
Finance sees cash and return. Operations sees completion and utilisation. Engineering sees technical integrity. Commercial sees revenue. HR sees people and policy. Technology sees systems. Safety sees risk. The airline, however, receives the combined consequence of all of them.
Published: 16 October 2026
Written by: Todd Skaggs
That distinction matters because aviation continues to operate with remarkably little financial room for error. IATA's June 2026 outlook expects the global airline industry to generate around $1.165 trillion in revenue this year but retain only $23 billion as net profit, equivalent to a margin of approximately 2%. Average net profit is expected to be about $4.50 for every passenger carried, while forecast return on invested capital remains below the industry's estimated cost of capital.
Aviation therefore does not need every department simply to become better at optimising itself. Senior leadership needs an organisation capable of recognising when individually sensible decisions are creating a poor enterprise result, challenging assumptions before reality does it for them, understanding how one choice travels through the rest of the business and acting while alternatives still exist.
That may become one of the defining leadership challenges of the next decade. Artificial intelligence, machine learning, optimisation and better enterprise data can help, but technology alone will not solve it. The deeper opportunity is to redesign how an aviation company makes decisions.
The Airline Exists Between the Departments
Consider a familiar operational problem. An aircraft becomes unavailable and Operations identifies the cancellation that best protects the remaining programme. From an aircraft-rotation perspective, the recommendation may be entirely rational, yet that flight could also be carrying connecting passengers, valuable cargo, a crew needed elsewhere tomorrow and an aircraft due at another station for maintenance.
Commercial may therefore prefer another cancellation. Crew Control may identify a different solution because of legality. Engineering may object because of maintenance positioning, while Finance may favour the option with the lowest immediate compensation and recovery cost.
The problem is not that one department is wrong. The problem is that no departmental answer necessarily represents the best answer for the airline.
This creates what can be described as an Enterprise Optimisation Gap: the distance between what is rational for an individual function and what produces the strongest result for the organisation as a whole. It exists during disruption, but it also appears in fleet planning, recruitment, technology investment, outsourcing, mergers, schedule design, maintenance, transformation programmes and capital allocation.
A finance team can reduce cost while unintentionally increasing disruption exposure. Operations can maximise aircraft utilisation while eliminating recovery margin. Commercial can open an attractive route whose complexity spreads cost elsewhere in the network, while HR can defend a perfectly reasonable remuneration structure as an important operational role remains vacant. Every department can improve its own scorecard while the enterprise becomes weaker.
Closing that gap is one of the few responsibilities that can only be performed by senior leadership. It requires leaders to stop asking merely whether a proposal works within the department presenting it and start asking what the decision does to the whole organisation over different time horizons.
Swissair Shows Why Strategy Needs an Organised Challenger
Swissair remains one of aviation's most important examples of what can happen when strategic ambition grows faster than an organisation's ability to challenge it. The original Hunter Strategy of the late 1990s contemplated comparatively modest minority investments in selected European airlines, with an initial identified capital requirement of approximately CHF300 million. What followed became much more aggressive, with SAirGroup spending roughly CHF4.1 billion in cash acquiring interests in carriers including LTU, AOM, Air Liberté, Air Littoral, Volare, LOT and South African Airways.
By October 2001, financial exposure associated with airline investments had reached approximately CHF5.9 billion. Subsequent investigation found significant weaknesses around financing, planning and crisis preparation, and on 2 October 2001 Swissair suspended operations as its liquidity crisis reached the point at which normal flying could no longer continue.
It is easy to look back and conclude that management simply made bad acquisitions, but that interpretation is not especially useful to today's executive. Swissair faced a genuine strategic problem because European aviation was consolidating and the carrier needed greater scale and market access. The more important question is how a strategy that began with one level of ambition was allowed to evolve into something dramatically larger without an equally powerful increase in challenge, funding discipline and contingency planning.
Every major aviation strategy therefore needs an organised challenger. The board should know who is responsible for arguing that the strategy may no longer work, what evidence would trigger reconsideration and which assumptions are carrying the greatest financial exposure. Challenging the plan must be institutionally legitimate rather than treated as disloyalty to the executives who created it.
This is particularly important because strategies often fail gradually before they fail visibly. Early warning signs may be distributed across Finance, Operations, Commercial, Engineering and the external market, with no single indicator serious enough to force action. Leadership's responsibility is to connect those weak signals before they become one very strong result.
Finnair Shows Why Assumptions Need Expiry Dates
Finnair provides the opposite kind of lesson because its management did not create the event that undermined the existing strategy. For years, Helsinki's northern geography supported an effective connecting model between Europe and North Asia, particularly Japan. Russia's invasion of Ukraine and the subsequent closure of Russian airspace to European airlines changed that advantage abruptly, forcing some Asian routes to become 10% to 40% longer and certain journeys to Japan to increase by as much as 40%.
Finnair recorded a comparable operating loss of €163.9 million in 2022. Management could have treated the closure as temporary and structured recovery around the hope that the old network economics would return, but instead it reduced parts of its North Asian operation, increased capacity towards India and North America, strengthened partnerships and used wet-leasing arrangements to deploy otherwise underutilised aircraft and employees more productively.
Most importantly, Finnair redesigned its strategy around the possibility that Russian airspace might remain unavailable. By 2023, comparable EBIT had reached €184 million, and the airline reported achieving its targeted comparable operating-profit level earlier than planned.
The lesson is bigger than geopolitics. Every aviation strategy contains assumptions that gradually stop feeling like assumptions because they are repeated so often: aircraft will arrive when scheduled, engine turnaround will remain within a particular range, fuel will stay within certain planning parameters, pilots will remain available, airport expansion will happen, a supplier will continue performing and market access will remain open.
Senior leadership should give material assumptions an expiry date. Every assumption capable of materially changing the business plan should have an owner, an evidence base, clear warning indicators and a point at which it must be actively revalidated rather than silently transferred into another forecast.
The board should not only ask, "What do we believe?" It should also ask, "When did we last prove that we should still believe it, and what will we do if it stops being true?"
Thomas Cook and Virgin Australia Show Why Leaders Need Another Move
Thomas Cook demonstrates why strategy ultimately depends upon optionality. The group generated almost £10 billion in turnover in 2018 and served around 22 million customers annually, yet entered compulsory liquidation in September 2019. Approximately 150,000 UK customers were abroad when the company collapsed, while thousands of employees lost their jobs.
Size had not preserved the company's ability to change direction. By the time financial pressure became critical, many conceivable strategic responses required something the business had increasingly little of: capital, time or both. Management can recognise that a business model needs to change and still discover that the organisation no longer possesses the resources required to travel from the old model to a new one.
Virgin Australia provides a different illustration. The airline entered voluntary administration during the pandemic in April 2020 and was ultimately acquired by Bain Capital, after which it changed its positioning, cost structure and operating model rather than attempting simply to reconstruct the business that had entered administration. By FY2023 it had reported its first statutory profit in 11 years, and by FY2025 revenue had reached A$5.81 billion with underlying EBIT of A$664 million before its return to the Australian Securities Exchange.
The broader leadership lesson is that resilient organisations preserve the right to make another move. Liquidity creates options, but so do alternative suppliers, fleet flexibility, workforce depth, available management capacity, adaptable technology and operational reserves. A business becomes strategically fragile when success depends upon the original plan continuing to work exactly as expected.
Senior executives should therefore add another question to conventional strategy reviews: if our principal plan became impossible tomorrow, how many credible alternatives would we still have? A large organisation with only one viable route forward may be considerably more fragile than its scale suggests.
Lufthansa and SWISS Show That Transformation Should Not Mean Making Everything the Same
Optionality also matters during mergers and transformation. When Lufthansa agreed to acquire SWISS in 2005, the integration did not require the Swiss carrier to surrender every distinctive element of its organisation. SWISS retained its brand, management, fleet, crews and Zurich hub while gaining access to Lufthansa's wider network, purchasing power, frequent-flyer infrastructure and multi-hub system.
The initial expectation was for annual synergies of approximately €160 million from 2007, yet reported benefits had already exceeded €200 million by 2006. The interesting point is not simply that the transaction generated value, but that integration was selective rather than indiscriminate.
This is relevant to almost every large transformation programme. Aviation companies often equate transformation with standardisation, centralisation and process consistency, yet the executive challenge should be deciding where commonality genuinely creates value and where difference should deliberately be preserved. Integration becomes dangerous when organisational tidiness is confused with economic improvement.
The same test should be applied to technology, shared services, acquisitions and operating-model changes. What should become common because duplication destroys value, and what should remain independent because autonomy, specialist expertise or local responsiveness creates value? Senior leadership should know the answer before the transformation begins.
Southwest and Delta Show Why Technology Recovery Is Not Enterprise Recovery
Southwest's December 2022 disruption exposed another weakness in traditional management thinking. Severe winter weather initiated the event, but the resulting scale of aircraft and crew displacement overwhelmed recovery processes and tools, contributing to more than 16,700 cancellations between 21 and 31 December. Southwest estimated the pre-tax financial impact at approximately $800 million, including around $410 million in lost operating revenue and roughly $390 million of additional net expense.
The airline subsequently strengthened crew-recovery capability, communications, network-health indicators and operational staffing, while budgeting more than $1.3 billion for information technology investment, upgrades and maintenance during 2023. The important lesson is not that Southwest lacked technology, but that the combination of technology, data, people and recovery processes encountered conditions beyond those they could absorb effectively.
Delta's experience during the July 2024 CrowdStrike outage reinforces the point. The problematic update itself was rolled back within approximately 78 minutes, yet Delta subsequently reported around 7,000 cancellations over five days affecting approximately 1.4 million customers. The airline attributed roughly $380 million of lost revenue and around $170 million of additional non-fuel costs to the disruption.
A technical event measured in minutes therefore became an operational recovery measured in days and a financial consequence measured in hundreds of millions of dollars. Once aircraft, crews, passengers, baggage and maintenance requirements have moved away from their planned positions, restoring the underlying technology does not restore the airline.
This should change how senior executives think about technology resilience. The important question is not simply whether a system has backup or how quickly the application itself can return. Leaders need to understand how quickly the organisation can reconstruct a common and reliable picture of reality once the operation has fragmented.
AI Could Finally Allow Leaders to See the Whole Airline
For decades, airlines have invested in sophisticated specialist systems. Crew applications understand legality, maintenance systems understand technical status, revenue-management platforms understand pricing and demand, customer systems understand passengers, while operations platforms understand aircraft rotations. The problem is that the airline itself exists in the interaction between all of them.
This is where artificial intelligence, machine learning and advanced optimisation become strategically interesting. Their greatest potential may not be automating another departmental activity, but creating an intelligence layer through which the organisation can understand the wider consequences of a decision before making it.
Cathay Pacific has been working with BCG and KLM on AI-supported transformation of its Integrated Operations Centre, using machine learning to predict delay risk and optimisation technology to evaluate recovery choices across aircraft, customers, crew, cargo and network performance. More than 15 teams and around 60 subject-matter experts have been involved, with reported benefits including millions of dollars of operating and disruption savings.
SWISS has pursued a related approach with an Operations Decision Support Suite developed with Google Cloud. Rather than replacing every specialist platform, the system brings passenger, crew, aircraft-rotation and technical information into a shared environment from which AI and machine learning can examine alternatives. Approximately half of SWISS network flights have been optimised using the approach, demonstrating how an intelligent layer can sit above existing systems rather than requiring the airline to replace everything underneath it.
LATAM shows how the principle can extend beyond disruption management. Working with McKinsey and QuantumBlack, it has deployed more than 50 reusable AI capabilities across areas including maintenance, cargo and digital customer journeys. Advanced planning and predictive maintenance tools reportedly released more than 1,400 additional aircraft days, an amount described as roughly equivalent to adding three aircraft without purchasing them.
That should change the executive conversation about AI. The question is not only how much artificial intelligence can reduce cost or how many routine tasks it can automate. A much more valuable question is how much capacity the company already owns but cannot currently access because information, decisions and departmental objectives are poorly connected.
Do Not Spend Millions Teaching Every Department to Optimise Its Own Silo
There is also an important danger. If every department develops AI independently, aviation companies could spend enormous amounts of money making their existing silos more intelligent without making the enterprise itself any smarter.
Revenue Management could become better at maximising revenue, Engineering better at predicting maintenance, Operations better at aircraft utilisation and Finance better at reducing cost. Senior leadership could then receive four increasingly sophisticated recommendations that remain incompatible with one another.
An airline could effectively spend millions on artificial intelligence and simply make every department better at disagreeing with every other department.
The larger opportunity is enterprise intelligence. Specialist systems should continue doing what they do well, but leadership needs a layer capable of identifying where one function's optimum creates another function's problem and showing the combined consequence before the organisation acts.
Air France-KLM's work with Accenture and Google Cloud illustrates part of that transition. The group has been developing a common environment in which generative, agentic and machine-learning models can be built, tested and governed across functions including engineering, ground operations and customer activities. The more important organisational principle is that AI becomes shared capability rather than a collection of disconnected experiments.
This is also where external consultants and technology specialists can help. BCG and BCG X, the KLM partnership, McKinsey and QuantumBlack, Accenture, Google Cloud and other specialist providers are already working on different aspects of airline optimisation and decision support. The senior executive's first question, however, should not be which vendor to appoint; it should be which decisions the organisation needs to make better.
Start With the Ten Decisions That Cross the Most Boundaries
A practical transformation does not need to begin with an enormous AI programme. Senior leadership could start by identifying perhaps ten recurring decisions whose consequences regularly cross multiple functions. They might include major cancellations, fleet allocation, schedule reductions, maintenance recovery, critical hiring, outsourcing, new-route approval, technology investment or the allocation of scarce training capacity.
Management should then map what information is required for each decision, which departments currently provide it, who recommends an action, who can challenge it and who ultimately has authority to decide. The exercise should also identify which costs or consequences are currently invisible because they fall into another department's budget.
This can expose an important weakness. Many organisations know their reporting lines in considerable detail but have never explicitly mapped how their most consequential decisions are supposed to travel through the business.
That difference matters because organisational charts show where people work. Decision maps show how the company actually moves.
Audit the Time Between Knowing and Acting
Aviation measures time with extraordinary precision. Block time, turnaround, taxiing, crew duty, maintenance intervals and departure delays are monitored constantly, yet senior management often knows much less about the time consumed inside its own decision process.
One useful measure is Decision Latency: the period between the point at which sufficient evidence exists that the existing course is no longer working and the point at which somebody with authority commits the organisation to another course. Delay may be entirely justified when safety, regulation or technical analysis requires further investigation, but it can also result from fragmented information, unclear accountability, repeated escalation or reluctance to abandon the original plan.
Executives can examine this retrospectively without buying any technology. Select several significant decisions from the previous year and reconstruct the timeline: when was the problem first visible, when was sufficient information available, when did ownership become clear, when was the choice authorised, when did implementation begin and what accumulated during the interval?
That analysis may uncover three different problems. Information can be slow to assemble, authority can be unclear, or execution can be slow after the decision has already been made. Treating all three as "slow decision-making" prevents management from fixing the actual obstacle.
AI may dramatically reduce information delay by bringing fragmented data together, but it cannot solve unclear authority. Faster information creates value only when somebody is empowered to act upon it.
Stress-Test the Leadership Team, Not Only the Operation
Aviation is unusually good at exercising emergencies. Airlines, airports and other aviation organisations rehearse situations because safety management recognises that nobody wants the first test of an emergency procedure to be the real emergency.
Senior decision-making deserves similar treatment.
An executive team could be presented with a realistic but hypothetical disruption: Russian airspace becomes unavailable, engine turnaround suddenly increases by 30%, a critical technology provider fails, a fleet subtype is grounded, a major base loses a significant group of captains or an airport unexpectedly reduces capacity for several days. New information can then be introduced progressively while management's response is observed.
The purpose would not be to determine whether executives can guess the "correct" answer. It would be to discover where information is missing, where two executives believe they own the same decision, where nobody owns it, which assumptions fail, how quickly financial consequences become visible and where the organisation discovers that its backup is not actually independent.
Banks stress-test capital. Aircraft manufacturers test systems, airlines test emergency response and cyber teams conduct attack simulations. Aviation leadership could similarly benefit from an executive decision stress test designed specifically to expose weaknesses before real events do it at commercial scale.
Move the Boardroom From Reporting to Simulation
Most senior dashboards remain dominated by what has already happened. They show punctuality, cancellations, revenue, staff turnover, maintenance performance, customer metrics and training results, all of which are essential but largely retrospective.
The more interesting future is a boardroom capable of testing the company before reality does. What happens to the summer programme if engine turnaround increases another 20%? What happens if three aircraft remain unavailable, if fuel rises sharply, if a major hub loses capacity for six hours or if 40 experienced employees are moved out of the operation to support a transformation programme?
Machine learning and advanced modelling can increasingly turn those questions into detailed scenarios. Instead of leadership debating one forecast, management can examine several possible futures and decide in advance what conditions would trigger a different response.
This changes the role of data from explaining the organisation to testing it. More importantly, it gives executives an opportunity to make difficult choices while the company still has time, cash and alternatives rather than after the constraint has already become operational.
Decide Which Resilience Is Worth Paying For
Better decision-making does not eliminate the tension between efficiency and resilience. A spare aircraft, reserve crew, second supplier, duplicate communications route, additional maintenance capacity or excess liquidity all appear expensive when they are not being used.
Their economics change when the primary route fails.
Leadership should therefore make resilience an explicit strategic choice rather than allowing it to be progressively removed through hundreds of individually sensible efficiency decisions. The question is not whether an airline should carry unlimited backup capacity, which would be economically unrealistic, but which failures would be sufficiently expensive that preserving another route is worth paying for.
Southwest's recovery difficulties, Delta's prolonged operational consequences, Finnair's geopolitical reset and Swissair's diminishing financial optionality all illustrate different versions of the same principle. An organisation rarely discovers the value of alternatives while everything is working normally.
The board should know where the business has deliberately accepted a single path and where it has deliberately paid to preserve another.
The CEO Does Not Need All the Answers
The traditional image of a successful aviation chief executive is someone with enough knowledge and experience to understand the business intuitively. Experience will remain enormously valuable, but the complexity of the modern aviation enterprise makes the idea that any individual can mentally model the whole organisation increasingly unrealistic.
A large airline may operate hundreds of aircraft, employ tens of thousands of people, process millions of itineraries, manage multiple fleets and bases, depend on global suppliers and connect thousands of operational decisions through increasingly complex technology. No individual can continuously understand every interaction.
The definition of strong leadership may therefore need to evolve. The exceptional executive will not necessarily be the person who personally has the best answer to every problem, but the person who builds an organisation capable of assembling the right evidence, revealing competing consequences and reaching a better answer than any individual function could produce alone.
That is where artificial intelligence could become genuinely transformational. Its highest value may be neither replacing employees nor automating decisions, but allowing leadership to see relationships across the enterprise that conventional organisational structures have historically hidden.
AI should not replace executive judgement. It should make the consequences of that judgement more visible before the decision is made.
Building a Better Aviation Decision System
Aviation history provides no formula guaranteeing success. Swissair demonstrates the danger of strategy outrunning challenge, Finnair shows the value of abandoning an assumption when reality changes, Thomas Cook demonstrates what happens when strategic options disappear, Virgin Australia shows how restructuring can create new ones, and Lufthansa's integration of SWISS shows that transformation does not require eliminating every difference.
Southwest and Delta demonstrate that recovering technology is not the same as recovering an airline, while Cathay Pacific, SWISS, LATAM and Air France-KLM show how advanced analytics and AI are beginning to connect information across traditional organisational boundaries. Taken together, those examples suggest that the next improvement in aviation leadership may come less from another individual management technique and more from redesigning how the enterprise itself reaches decisions.
For a senior leadership team, that means identifying the decisions that cross the greatest number of organisational boundaries, putting expiry dates on the assumptions carrying the most strategic exposure, measuring the delay between knowing and acting, stress-testing executive decision processes and deciding explicitly where resilience deserves investment. AI and machine learning should then be directed towards those problems where better cross-functional intelligence can materially improve the quality of the choice.
The competitive advantage may not belong to the airline with the most artificial intelligence, the largest technology budget or the greatest quantity of data. It may belong to the organisation that knows which decisions deserve better intelligence, which assumptions must continually be challenged, which departmental objectives cannot all be maximised simultaneously and which decisions must remain accountable to human judgement.
Aviation leadership has never required predicting every war, bankruptcy, technological failure, pandemic or operational shock. It requires building an organisation capable of recognising when reality has diverged from the plan, understanding what that change means across the enterprise and acting before the number of available options has narrowed to one.
That is the opportunity for the next generation of aviation leadership: not to know everything, but to build a company that can see enough, soon enough, to make a better decision while there is still time to make one.
Key Facts
Departments can make individually rational decisions that collectively produce a poor enterprise outcome.
The article introduces the concept of the Enterprise Optimisation Gap, the distance between local optimisation and the best overall outcome for the organisation.
Airline leadership increasingly depends on connecting decisions across operations, finance, commercial, engineering, HR, technology and safety functions.
Swissair demonstrates the dangers of strategy expanding faster than challenge, governance and contingency planning.
Finnair illustrates the importance of regularly reassessing strategic assumptions when external conditions change.
Thomas Cook and Virgin Australia show how organisational resilience depends on maintaining strategic optionality.
Southwest and Delta highlight how technology recovery and operational recovery are not the same challenge.
AI may create its greatest value by helping leadership understand enterprise-wide consequences rather than simply automating departmental activities.
Decision-making quality often depends on reducing "Decision Latency", the gap between recognising a problem and acting on it.
The future competitive advantage may belong to organisations that improve cross-functional decision-making rather than simply adopting more technology.
Related Articles
Planning growth, fleet changes or seasonal operations in 2026? Contact Brookfield to discuss your staffing and consultancy needs. Email: info@brookfieldav.com
Explore our full range of recruitment services, connecting aviation businesses with skilled pilots, aircraft engineers and industry professionals worldwide.
Author: Todd Skaggs Aviation staffing and consultancy insights LinkedIn
