The matching algorithm had been developed over twenty years of outcome data. The premise was simple and the results were consistent: the knowledge most needed by a person was almost always in the possession of someone who had wronged them or whom they had wronged. The algorithm identified the match. The organization facilitated the arrangement.
Tian worked in matching. She reviewed the algorithm's outputs and made the determinations that required judgment: cases where the wrong was too severe to constitute a viable teaching relationship, cases where the power differential was too steep, cases where the match was technically correct but practically impossible. The algorithm was good. Tian's job was the cases where good wasn't enough.
She had placed 847 matches in eleven years. The success rate, defined as the student completing at least three sessions, was 61%. The success rate defined as the student acquiring what they came for was harder to measure. Tian estimated it was lower. She did not tell the organization this.
Her own match had arrived three months ago. A sealed file with her ID and a name she had not seen in twelve years. She had put it in the bottom drawer of her desk.
She knew what the file would contain before she opened it. She had worked in matching for long enough to know her own case. The knowledge she needed, the specific gap in what she understood, was in the possession of one person. She had known this for years. She had built a career around helping other people open their files.
The drawer was locked. The key was on her keychain.
She made three placements and went to lunch.