Chapter 76: The Pattern, Case by Case
Samuel kept the word pattern gray while each matching box acquired a different class of proof.
On Thursday, Jonah's and Lila's columns filled the evidence matrix in KRR's review room.
Their departments, supervisors, dates, and underlying complaints were different.
The repeated label could not carry the conclusion by itself.
Samuel divided the proposed process into five testable nodes.
The first was a formal complaint or protected issue entering an authorized channel.
KRR required a timestamped source record rather than a witness's later memory that something had been reported.
Jonah's issue register and gateway record supported his entry.
Lila's native complaint and server receipt supported hers.
Neither record proved that the underlying concern was fully correct.
The second node was assignment of a reliability or cooperation risk label.
That node required the native case field and deployment history, not matching language copied into a summary.
D06 defined the routing field, and L05 showed the version deployed before both events.
The separately held case records placed the field after each protected issue within the tested timeline.
That sequence did not establish why a person chose the label.
The third node was transfer into the strategy-office reference workflow.
Asha's W02 statement described that routing only where she had personally handled it.
Lila's protected W03 statement described the corresponding instruction in her own case.
The two witnesses had not received each other's answers.
Their separate accounts matched the deployed field without becoming a population estimate.
The fourth node was a reference stabilization task tied to the marked record.
KRR limited the match to task ID, service category, and sequence while private questions stayed with their custodians.
The task identifiers appeared in separate custodian-controlled indexes, with all private contents still restricted.
KRR could test identifiers and timestamps without giving me either person's reference questions.
Samuel marked the first four boxes SUPPORTED FOR TWO TESTED CASES.
Grant objected that the word pattern implied a company-wide practice.
Samuel kept the word gray and added the qualifier tested process correspondence.
Vivian's attorney argued that different managers broke any meaningful connection.
The matrix showed why different managers mattered.
The same deployed fields and route appeared despite the departmental separation.
That fact strengthened a process question but did not answer motive.
The fifth node concerned external professional harm caused by the contact.
That final node required an outside decision-maker's record or testimony capable of separating the contact from other reasons.
Jonah had documented lost opportunities, but the independent decision records remained incomplete.
Lila had not authorized KRR to obtain every outside decision file.
Neither witness could speak for an employer's or client's full reason.
Samuel deliberately left the fifth box completely white.
I asked whether four supported nodes were enough to call the protocol used against complainants.
He limited the answer to the two records before us.
The evidence supported complaint entry followed by risk routing and reference-task activation.
It did not yet support a final causal finding about outside loss.
KRR updated the core status without changing that open field.
The matrix listed each source by custodian, format, and access level.
The witness statements remained separated from company records and from each other.
The unique meeting recording carried no pattern weight.
It belonged to Andrew's earlier language, not these cross-case system nodes.
Samuel locked the two-case comparison and opened the queue for the remaining records.
The identity custodian returned only match-status notices, not names.
One notice concerned an operations record already associated with Devon's protected tasks.
The next concerned a project-management record whose witness consent remained pending.
KRR did not yet possess the underlying access log for that third case.
The custodian's restricted status report nevertheless showed the same four-node shape.
Its protected witness was Malik, and he had not decided whether to be named.