By Thursday evening, the reviewers had confirmed multiple weaknesses in the system's handling of interpreter requests and multilingual consent wording. They had found no basis to declare that every translated interaction had caused harm. They had also found no reliable process to detect or correct errors before those errors affected decisions. The distinction mattered. Marian insisted it remain clear in their report.
The launch was scheduled for the following morning in the ballroom of a downtown hotel. Invitations had already gone out to reporters, donors, hospital administrators, and local officials. Adrian argued that canceling would make the foundation look incompetent. The public-relations team proposed a short statement about "minor refinements" while the program proceeded with a limited release.
Alistair read the proposal and put it facedown.
"No," he said. "If I knew what we know now and let people think it was ready, that wouldn't be a refinement. It would be a lie."
The following morning, Marian arrived at the hotel carrying a folder and wearing the navy suit she had purchased for her first professional conference. Near the entrance, she saw an arrangement of white flowers and an enormous screen displaying the words OPEN DOOR: A FUTURE WITHOUT BARRIERS. The sight made her stomach tighten. An elegant slogan could make the truth seem inconvenient.
Alistair stepped onto the platform at precisely ten. He did not introduce a product demonstration. Instead, he told the audience that the foundation was suspending the planned rollout of its translation platform's higher-risk features while independent language-access and patient-safety reviews were conducted. Routine information tools would be evaluated separately. The clinics would continue receiving funding for qualified human interpreters, and no site would be pressured to rely on technology it did not trust.
"The problems were brought to our attention by staff and community members who refused to accept a reassuring presentation as proof," Alistair said. "I should have asked better questions before promising a launch date. I take responsibility for that failure."
He did not name Marian as the person who had saved the initiative. She had asked him not to. Instead, he named the people who would help redesign it: nurses, patient advocates, qualified interpreters, clinicians, and independent reviewers.
During questions, a reporter demanded to know whether Rowan Technologies had knowingly ignored safety warnings. Alistair answered that the review would examine the full decision-making record, including his own role, and that findings would be made public with necessary privacy protections. It was less satisfying than a dramatic accusation, but it was the truth he could support.
By lunchtime, the story had spread across business news outlets. Some commentators praised the decision. Others called it an expensive embarrassment. One headline described the failed launch as proof that the billionaire's philanthropy was simply a marketing strategy. Another claimed, incorrectly, that all translation technology was unsafe. Marian spent much of the afternoon correcting that misconception with the communications team.
"We can't replace one exaggeration with another," she told them. "This isn't about convincing people computers are useless. It's about knowing when a person needs another person."