What If an Airline Had Process Mining? (They'd have known why in days, not weeks)

Martin Barclay — Mon, 23 Feb 2026 22:21:49 GMT

A quick note before we get into it: An airline boarding process is inherently physical — people walking down aisles, stowing bags, finding seats. You can't fully digitize that into a workflow with all the data points needed to mine it holistically. But if it WERE possible — the principle behind what happened here maps directly to what we see every day with ServiceNow workflows that don't yet take advantage of the insights-to-action that Process Mining offers. And that's what makes this story worth telling.

A major airline just did something it hadn't done in over 50 years. In January, the carrier scrapped its longstanding open seating policy and moved to assigned seats, a new boarding group system, and a WILMA boarding order (window seats first, then middle, then aisle). It was a massive process change — one the airline had spent months planning, running computer simulations, and testing with in-person walkthroughs.

And it took nearly a month after it launched the new process — weeks of widespread customer complaints, social media backlash, and crew frustration — before the airline had a clear enough picture of the problems to start acting on them.

Passengers in premium extra-legroom seats near the front of the cabin were boarding only to find that all the overhead bin space above their rows had already been taken by earlier boarding groups seated farther back. They'd then have to swim upstream through the aisle to stow their bags in the back of the plane, then fight their way forward again to sit down. Passengers stuck in middle seats on half-empty flights were being told they couldn't move to empty window seats — even in their own row. One travel writer described the whole scene as "far more chaotic than the other airlines."

To their credit, the airline says it was watching from the start.

A spokesperson told the press the airline has been "closely monitoring input and real-world behaviors to validate our assumptions and identify where we can refine the experience." But despite that monitoring, it still took nearly a month to translate observations into announced adjustments — restructuring boarding groups, moving flight attendant bin storage to the back of the plane, shifting premium and elite passengers to earlier boarding positions. By then, the damage was already done. Weeks of bad headlines, frustrated loyal customers, and a growing narrative that the airline had botched the rollout.

Now, you might be wondering what any of this has to do with Process Mining.

The answer: almost everything.

What the airline got right (eventually)

What this airline is doing right now — observing real behavior, comparing it to their planned process, identifying where things break down, and making targeted adjustments — is essentially process mining in action. They just didn't have the advantage of doing it systematically, with data, at scale, and within days of launch instead of weeks.

Process Mining is a data analysis approach that lets you see how work actually flows inside your organization. Not how you think it flows. Not how it was designed to flow. How it actually flows, based on the digital footprint left behind in your systems. Think of it as an x-ray for your business processes — it shows you the reality of how work gets done, where it stalls, where it takes unexpected detours, and where there are opportunities to improve.

This airline essentially did this the hard way. They launched a major process change, assumed their models would hold, and then had to scramble when real-world human behavior didn't cooperate. Passengers didn't follow the predicted patterns. The interaction between the new boarding groups, the WILMA system, recently introduced checked bag fees (which put more carry-ons in overhead bins), and passengers' natural instincts created friction that no simulation fully anticipated.

Sound familiar? It should. Every organization that runs complex workflows deals with a version of this. You design a process. You roll it out. And then people do what people do — they find workarounds, create bottlenecks, and behave in ways that the design didn't account for.

Evidence in Days, Not Complaints Over Weeks

Here's where this story really connects to what we think about every day on the Process Mining team at ServiceNow.

The airline had to rely on customer complaints, social media posts, crew feedback, and anecdotal observation to figure out what was going wrong. That takes time. It introduces bias. And it means you're reacting to problems after they've already frustrated your customers.

Organizations running workflows on ServiceNow don't have to wait. With ServiceNow Platform Analytics, workflow stakeholders can monitor KPIs from the moment a process goes live — detecting anomalies and performance issues immediately, with no delay at all. If something starts trending in the wrong direction, you know right away.

And once you've accumulated a representative sample of process data — a sufficient number of records, perhaps a couple of days' worth — you can run a Process Mining project to see exactly why things are going sideways. The variant paths, the rework loops, the bottlenecks, the deviations from your intended process. All based on evidence, not anecdotes. Platform Analytics tells you what changed. Process Mining tells you why. And because both sit on the same platform where the work runs, you can go from insight to action right away, without the integration overhead and handoffs that standalone tools require.

That's the difference between waiting for a flood of complaints and proactively analyzing actual process data while the issues are still fresh.

Process Mining shows how work actually flows in the real world

The Gap Between How You Designed It and How It Actually Runs

The airline spent months simulating their new boarding process. They ran models. They tested scenarios. And they still got surprised by reality. That's not a knock on the airline — it's a reminder that no model, no matter how thorough, can fully predict how a complex process will behave once real people start interacting with it at scale.

This is exactly why process mining exists. Not to replace planning, but to close the gap between how you designed a process and how it actually runs. To give you evidence-based visibility — on a regular cadence, not just when someone complains — so you can keep refining.

This airline is learning this lesson in public, in front of millions of passengers and a very vocal internet. Organizations running workflows on ServiceNow have the advantage of learning it with data — systematically, repeatedly, and with a direct path from the insight to the fix.

If this airline could have mined their boarding process digitally, Platform Analytics would have flagged the performance issues from day one. And within a couple of days, a Process Mining project would have shown them exactly why premium passengers in the front rows were boarding after the overhead bins were already full — and where to make adjustments. Well before it became a headline.

That's what our customers do every day with Process Mining.

Want to see how Process Mining works on your instance with your data? Here are a couple of resources to get you started:

post What If an Airline Had Process Mining? (They'd have known why in days, not weeks) in Process Mining blog

What If an Airline Had Process Mining? (They'd have known why in days, not weeks)

What the airline got right (eventually)

Evidence in Days, Not Complaints Over Weeks

The Gap Between How You Designed It and How It Actually Runs