There she met Naomi Bell and a forensic accountant named Felix Mercer. Felix had been hired to examine inconsistencies in maintenance expenditures, an assignment that initially looked too modest to trouble the board. He placed three invoices on the table and drew a line through the vendor names.
“They don't match,” Elena said.
“They're not supposed to. Look at the routing numbers.”
Two companies in different jurisdictions had submitted bills for the same turbine inspection, each claiming specialist labor carried out on the same afternoon. The inspection itself had been logged by a third firm. A photograph attached to one invoice showed the wrong engine serial number. The error was small enough to escape a glance and clear enough to trouble anyone who looked closely.
Felix did not claim that every duplicate invoice was fraudulent. Clerical errors could produce remarkable coincidences. But the payments had been approved through an unusual authorization chain and linked to a maintenance-release override. The same override had appeared in the message Grant displayed on his phone.
Naomi warned them against collapsing the questions into a single theory. A passenger-access breach could coexist with financial misconduct without being caused by it. Their duty was to establish each link rather than construct the most satisfying narrative. Elena appreciated the caution even while she disliked the uncertainty.
Felix encountered the first significant disagreement within the investigation when two experts interpreted the same payment ledger differently. One believed an intermediary had deliberately split invoices to avoid financial approval limits. The other argued that the division could reflect separate categories of work under the original contract. Both explanations were plausible from the summary data.
Naomi asked the experts to prepare a joint statement identifying facts they agreed on and evidence required to distinguish the interpretations. They obtained the original purchase orders, shipping records, and acceptance certificates. Several transactions could be matched to documented services; others could not. The experts revised their initial positions rather than treating consistency as a form of loyalty.
Elena attended a briefing at which the differences were explained. She admitted that she had expected a clearer answer after so many months. Felix replied that reliable investigation often made the picture more complicated before it made it useful. The purpose of expert work was not to fulfill management's expectations, but to show which conclusions the documents could bear.
An independent auditor later discovered that one suspicious approval had been entered by a temporary finance worker who was covering an absent colleague. The worker had followed a written instruction and had no authority to verify the technical scope of the expenditure. The investigation cleared her of misconduct and recommended changes to the approval design. Elena insisted that the correction be communicated directly to the employee before any public summary was issued. A person's reputation should not wait for the communications timetable.
The unresolved transactions were referred to the appropriate authorities and pursued through contractual recovery where justified. Felix maintained separate files for established facts, open questions, and matters subject to legal proceedings. That discipline prevented the board from mixing likely loss estimates with confirmed findings.
During a late meeting Miriam asked whether the company would ever be able to say the matter was closed. Felix answered that specific actions could be completed, payments recovered, contracts revised, and systems tested. Some questions might remain unresolved. A company should not leave every issue open forever, but it should not declare certainty simply to make the quarterly report look tidy.
Elena thought about how often she had once asked teams for a single final number. The world had offered her estimates, qualifications, and incomplete records; she had preferred confidence. Now she asked for ranges and assumptions alongside the figures. She found the resulting decisions more difficult to explain and easier to defend.
When Felix delivered the revised report, he included a short appendix titled 'What Would Change These Conclusions.' Elena asked that the appendix remain in the board packet. It was the first time she could remember wanting directors to read a list of reasons they might be wrong.
They traced the approval history. One signature belonged to a director who had retired six months earlier. Another was attributed to a staff engineer who had been on medical leave on the date in question. His electronic credentials had nonetheless been used to sign two releases. The company had disabled those credentials only after his leave ended, leaving an unexplained interval in which another person might have used them.