By 10:18 a.m., the investigation had expanded beyond Flight 728.
Natalie Price and Peter Lawson had moved into a larger operations room, where several monitors displayed system records, passenger service logs, and the status of preservation requests. A compliance analyst named Rachel Kim joined them with a preliminary summary of complaints submitted to the airline over the previous eighteen months.
The summary was not a verdict.
It was a starting point.
Rachel explained that the airline had received hundreds of complaints involving boarding disputes, identification checks, seat changes, and requests for passengers to leave aircraft. Most had been resolved through ordinary customer service procedures. Some had involved legitimate safety concerns or reservation errors.
But a smaller group shared several characteristics.
The passengers had been described as suspicious or noncompliant before the original concern was clearly identified. The initial reports frequently relied on vague language such as does not appear to belong, unusual presentation, or possible premium-cabin access concern. In several cases, later reviews had found no ticketing discrepancy.
The reports were not identical, and the data did not establish a single cause.
Still, the similarities were troubling.
"How many cases are we talking about?" Natalie asked.
"Twenty-seven that warrant closer review," Rachel replied. "That's out of the larger complaint set. We have not yet determined whether all twenty-seven are comparable."
Peter leaned toward the screen.
"How many resulted in passengers being removed?"
"Nine. Four involved passengers who later received confirmation that their tickets had been valid. In two cases, the records are incomplete. The remaining three require additional review."
Natalie looked at the list.
"Were the decisions made by the same employees?"
"Not all of them. But there are recurring names among the people who initiated manual verification requests."
"Show me."
Rachel opened another report.
Angela Morris appeared in several records. Brian Keller appeared in others. Captain Richard Hayes was mentioned in two previous complaints, although neither had resulted in a formal disciplinary finding.
Natalie frowned.
"Did anyone examine the relationship between these cases before?"
"Some complaints were reviewed individually. I haven't found evidence of a consolidated analysis."
"Why not?"
"The reporting system categorizes them differently. A passenger removal may be recorded as an operational incident, while a manual identification check is classified as a customer service issue. Complaints about crew conduct go through another process."
Peter stared at the screen.
"So nobody was looking at the whole sequence."
"Not consistently," Rachel replied.
Natalie understood the significance immediately.
A system could appear effective when each department reviewed its own records. The boarding team would examine the verification request. The flight crew would document the removal. Customer service would respond to the passenger's complaint. Human resources might address an employee's conduct.
Each department could close its own case without recognizing that the same underlying assumptions had appeared repeatedly.
"Can we establish whether the passengers shared any demographic characteristics?" Natalie asked.
Rachel nodded.
"We can examine the available data, subject to the applicable privacy and review requirements. But we need to be careful not to infer protected characteristics from names or photographs alone. We should use reliable information where available and document the limitations."
"Agreed. I want the analysis to be methodologically sound."
Peter pointed to another column.
"Look at the reason codes. Several reports use the same generic phrase, even when the underlying circumstances differ."
"Who selected those codes?"
"Sometimes the gate agent. Sometimes a supervisor. Sometimes the crew member completing the incident report."
"Then we need to determine whether the codes are being used accurately or simply copied from earlier entries."
Rachel opened a sample case.
The record described a woman who had arrived at the gate wearing ordinary travel clothes. An employee had questioned her boarding group, although the reservation system showed a valid premium-cabin ticket. The woman had asked why she was being questioned, and the interaction had been recorded as a possible refusal to cooperate.
A later customer service review had confirmed that the ticket was valid.
But the initial incident report remained unchanged.
In another case, an older passenger had been asked to produce additional documentation after an employee questioned whether the passenger's appearance matched the expected profile of a premium traveler. The passenger had complied, but the crew had recorded the interaction as a security-related concern.
Again, the original report had not been corrected.
Natalie read both cases in silence.
"These reports don't prove intentional discrimination," she said finally. "But they raise a serious question about how subjective judgments are being converted into official records."
Rachel nodded.
"That's what I thought."
"Preserve the original entries. We need to know what the employees wrote at the time, not only what the records say now."
Peter added, "We should also examine whether supervisors instructed employees to use particular language."
"Yes. Search the training materials, internal guidance, and relevant communications. But keep the review within its authorized scope."
Rachel began making notes.
Across the airport, Evelyn Carter received a summary of the preliminary findings.
She read it slowly.
The pattern was not yet conclusive, but the problem was becoming clearer. A person could be singled out because an employee believed they did not fit an expected image of wealth or status. That suspicion could then be entered into an official system, where subsequent employees might treat it as verified information.
If the person questioned the decision, the questioning itself could be recorded as noncompliance.
The process was self-reinforcing.
The more a passenger challenged an unsupported suspicion, the more evidence the system appeared to generate against them.
Evelyn wrote three questions in her notebook.
Who defines what a suspicious passenger looks like?
What evidence is required before an employee initiates a manual verification request?
What safeguards prevent an unsupported concern from becoming an official allegation?
She sent Natalie a message requesting that those questions be included in the review.
Then she asked for something more.
"I want to know how many complaints were closed without the passenger receiving a clear explanation of the findings."
Natalie replied that the team would examine the issue.
Evelyn put her phone away.
She was no longer looking at a single confrontation. She was looking at a process that might have allowed dozens of people to be treated unfairly while giving the organization a misleading impression that every incident had been handled properly.
That did not mean every employee involved had acted with discriminatory intent.
It did mean the airline had a responsibility to examine whether its procedures encouraged unsupported assumptions and failed to correct them.
At 10:36 a.m., Richard Hayes received a formal request to participate in an independent interview after the flight, once operational duties permitted.
The notice made no finding of wrongdoing.
It listed the areas the review would examine: the original identification concern, the decision to order the passenger off the aircraft, the threat of arrest, and the captain's handling of the situation after the passenger presented her official credentials.
Richard read the notice twice.
Then he looked toward Thomas.
"Have you ever seen a company investigate one of its own captains this thoroughly over a passenger complaint?"
Thomas considered the question.
"I've seen investigations become thorough when the evidence warrants it."
Richard looked down at the notice.
For the first time, he began to understand that the problem might not be solved by convincing management that he had made a single poor decision.
If the airline had failed to recognize a recurring pattern, his conduct might be one part of a larger institutional failure.
That possibility frightened him.
It also presented a question he had spent his career avoiding.
If he wanted the investigation to be fair, he would have to accept the possibility that the evidence might not support his preferred explanation.
And if he wanted to remain a professional worthy of the uniform he wore, he would have to decide whether he was prepared to tell the truth even when the truth made him look bad.