Latest / The Operator's Manual with Fexingo: Daily Decisions, Tactical Wins, and Business Execution / How One Airline Mastered Real-Time Rebooking
Transcript
- Lucas: There's a moment in every airline operations center that separates the great from the merely competent. It happens about twenty minutes after a disruption hits, when the initial shock wears off and the team has to decide: do we fix this one flight at a time, or do we re-plan the entire network? Luna: And most airlines, historically, pick option A — flight by flight — because it feels safer. Less scope, less risk. Lucas: Exactly. But the cost of that incremental approach is staggering. I want to look at a specific case: a major US carrier — I'll call it Airline X — that got caught in a severe thunderstorm at its largest hub in June 2022. The storm lasted about three hours, but the ripple effects stranded over 50,000 passengers and took the airline seven days to fully recover. Luna: Seven days. That's not just operational failure — that's a brand crisis. I remember the headlines about luggage piles and sleeping bags on terminal floors. Lucas: Right. And what's interesting is what Airline X did after that storm. They didn't just blame the weather and move on. They spent the next eighteen months rebuilding their entire irregular operations playbook — they call it IROPS internally. And by the end of 2024, when a similar storm hit the same hub, they recovered in under 90 minutes. Luna: Ninety minutes versus seven days. That's not incremental improvement — that's a completely different operating model. What changed? Lucas: Three specific things. First, they built a real-time disruption dashboard that aggregates data from crew scheduling, maintenance, gate agents, and air traffic control into a single view. Before, each department had its own spreadsheet and they'd conference-call to reconcile. Now the system flags conflicts automatically. Luna: So instead of a bunch of people on a bridge arguing about whose data is right, everyone sees the same picture. That alone probably cut hours off the response time. Lucas: It did. But the second piece is where it gets really interesting. They deployed a crew-recovery algorithm that solves for the entire network simultaneously. When a flight cancels, the system doesn't just look for a replacement crew at that airport — it looks at every open seat on every flight in the next 24 hours and computes the optimal reassignment, including deadheading options. Luna: And this is where a lot of airlines screw up, right? They solve locally and create problems somewhere else. You reassign a pilot from a later flight to cover a cancellation, and then that later flight gets delayed, and the cascade continues. Lucas: Exactly. The old approach was a game of whack a mole. The new algorithm essentially runs a giant optimization problem — it weighs crew legality, costs, and customer impact — and produces a plan in about 30 seconds. The human ops team then reviews and tweaks, but the heavy lifting is done by the machine. Luna: What about the passengers? Because the third piece must be customer-facing, right? Rebooking 50,000 people manually is impossible. Lucas: That's the third piece: a passenger rebooking engine that's integrated directly into the airline's app and website. When a flight cancels, the system automatically checks the passenger's itinerary, their loyalty status, and their travel preferences — window seat, aisle, same-day arrival, whatever — and offers up to three alternative options instantly. No wait time, no call center. Luna: And customers actually use it? I've seen airlines roll out automated rebooking and passengers still flood the gate because they don't trust the algorithm. Lucas: They had a trust problem at first. But they added a feature: if you accept the automated option, you get a digital voucher for a meal or lounge access, depending on the delay length. The acceptance rate went from about 20 percent to over 70 percent within six months. People will trust a machine if there's a clear incentive. Luna: That's clever. It's basically a nudge — you're rewarding the behavior you want, which is keeping passengers out of the terminal queues so the agents can focus on the complex cases. Lucas: Exactly. And the result was measurable. In the 2024 storm, Airline X rebooked 85 percent of disrupted passengers within 90 minutes. The remaining 15 percent — the ones with connecting flights on other airlines, or group bookings, or pets — those went to human agents. That's a huge shift from 2022, where every single passenger was a manual case. Luna: Let's talk about the cost side. Because building a custom algorithm and rolling out a new dashboard isn't cheap. Do we know what they spent? Lucas: The airline didn't disclose the exact figure, but industry analysts estimate the total investment in the IROPS transformation was somewhere between 50 and 80 million dollars over two years. That includes software development, hardware, training, and change management. Luna: And the payoff? Aside from the obvious reputational benefit. Lucas: The airline itself calculated that the 2022 storm cost them roughly 200 million dollars in direct operational costs — refunds, hotels, meal vouchers, overtime, plus lost revenue from passengers who switched airlines afterward. They also saw a measurable dip in load factor for the next three quarters, which they attribute to the reputational damage. So the ROI on the 50 to 80 million was actually quite fast — basically one major storm avoided pays for the whole system. Luna: Which is the real lesson here: it's not about perfect weather forecasting. It's about having a system that can absorb a shock and keep functioning. That's true for any business, not just airlines. Lucas: Right. And that brings up a question I think about a lot: why do so many companies underinvest in ops resilience? They know disruptions will happen — they've seen it happen to competitors — but they still treat it as a cost center rather than a competitive advantage. Luna: I think it's because resilience is hard to measure until the moment you need it. You can't put a line item on a balance sheet that says 'chaos avoided.' And most CFOs are trained to optimize for the steady state, not the edge case. Lucas: That's a great point. And it's also why I love this show. We get to dig into the operational details that most people don't see. If you're listening and these conversations have helped you think differently about your own work — even in a small way — we want to mention that listener support is what keeps this podcast independent and ad-free. Luna: Yeah, no sponsors, no corporate overlords. Just us and a microphone, digging into operations stories like this one. Lucas: If that's something you'd like to help sustain, you can find us at buy me a coffee dot com slash fexingo. No pressure, truly — just a way for listeners who get value from the show to chip in. And we mean it when we say it helps us keep doing this. Luna: Alright, back to the ops room. So Airline X solved the acute disruption problem. But what about the chronic stuff? The everyday delays that add up? Lucas: That's actually where they're focusing next. They built a predictive model that flags flights at risk of delay before the schedule is even published — based on weather forecasts, air traffic flow, crew availability, and maintenance history. The idea is to proactively adjust the schedule a day in advance rather than reacting when the delay happens. Luna: So they're moving from real-time response to proactive prevention. That's a whole other level of operational maturity. Lucas: It is. And the early results are promising. In the first quarter of 2026, Airline X saw a 15 percent reduction in departure delays that were classified as 'controllable' — meaning delays caused by crew or maintenance issues, not weather or air traffic control. Luna: That's a big number for something that doesn't require a single new plane. It's pure process improvement. Lucas: Exactly. And it shows that the next frontier in airline operations isn't about buying more airplanes or hiring more agents. It's about using data to make better decisions faster. The same principle applies across industries — whether you're running a supply chain, a hospital, or a software company. Luna: So what's the one takeaway for a business leader listening? If they can only implement one piece of this playbook, what should it be? Lucas: I'd say: build a single source of truth for your operations data. The airline's biggest bottleneck before the transformation wasn't the weather — it was that no one had a clear picture of what was happening across the network. Once they fixed that, everything else got easier. Luna: Integration before automation. That's a good rule to remember. Lucas: Absolutely. Get the data flowing first, then worry about the algorithms.