Marla Bennett arrived a little after seven carrying a binder with a worn blue spine. She had gone home briefly to see her husband and leave a plate of food for their relatives, then returned because she did not want the board to discuss the matter without the old technical records. She apologized for the size of the binder. Audrey told her not to apologize for bringing more evidence than a convenient summary could hold.
Ellis joined them with another auditor. They spread three years of reports across the conference table, sorting them by date and marking which documents were observations, which were formal requests, and which had been accepted for review. At first the papers looked unrelated: a controller fault in Indiana, a laboratory recommendation, an engineering note about unusual loads, and a rejected budget for an additional test environment. Gradually, the common pattern became visible.
The same approval method had been used across successive product generations. It relied on a standard set of tests that usually demonstrated satisfactory performance but did not reliably reproduce the unusual load combinations where the feedback delay appeared. Nolan's reports had not alleged that every unit would fail. They had warned that the evidence used to certify certain performance claims was incomplete, and that the testing procedure needed to cover conditions the equipment could plausibly encounter in service.
Marla placed a three-year-old inspection report on the table. Along the bottom was her handwriting: "Escalated to Blackwell. No corrective action approved." She had written the note after a meeting in which a budget request for expanded testing was deferred. The note had seemed routine at the time, part of the frustrating administrative work that accompanied technical management. Tonight it looked like a marker on a road no one had bothered to follow.
"Why didn't this reach me?" Audrey asked.
Marla did not answer immediately. "Some of it probably did, in condensed form. The quarterly briefing described the concern as low priority and said engineering was monitoring it. That wording wasn't technically false when the first report was written. But the monitoring never produced the additional testing that would have let us resolve it."
Audrey turned through the pages. "And the later reports?"
"They were acknowledged. I asked for a decision several times. The budget always came back deferred."
"Why didn't you come straight to the executive office?"
Marla looked at her for a long moment. "I did once, through the formal escalation route. The response came from Operations saying the matter had been reviewed. After that, I was told repeated escalation would be treated as a failure to follow the chain of command."
The sentence settled heavily between them. Audrey had insisted on formal channels because they produced a record, yet those same channels had become barriers when the people controlling them wished to avoid bad news. She thought about the annual employee surveys reporting concerns about speaking up. Her leadership team had described the responses as ordinary resistance to change. She had accepted the explanation because the financial performance looked strong.
"I should have seen the pattern," she said.
"It's easier to see it from down here," Marla replied, then looked embarrassed at the phrase.
"Don't apologize. I asked the question."
Ellis found the original budget proposals and compared their approval histories with the old records. Each had reached an office controlled by the chief operating officer before being postponed or closed. That did not prove a malicious intention behind every decision, and he wrote the limitation plainly. Together, however, the records established that management had received technical warnings and had not authorized the verification work needed to resolve them.