The investigation finding I got wrong

Situation. A team lead on one of our sites was accused of repeatedly mocking a colleague’s accent. I interviewed the person who raised it, the team lead and one witness. The complainant’s account was detailed, consistent and specific about dates. The team lead denied it. The witness said they “might have heard something.” I drafted the finding as substantiated.

Action. Before it went out, my manager read the draft and asked one question: what corroborates it? Honestly, nothing did, apart from how convincing the interview had felt. We changed the finding to inconclusive. We kept the full record, arranged a facilitated conversation, and I booked a check-in with the complainant for four weeks later. I explained to the complainant, in person, that inconclusive meant we could not prove it either way, not that we thought they were lying. That conversation was harder than the investigation.

Observation. Two months later, a second complaint about the same team lead came from a different shift, with a message that matched the first account closely. With the first record on file, the pattern was clear and we could act. If I had substantiated the first complaint on belief alone, I think the team lead’s appeal would have succeeded, and the second person would have walked into a process that had already lost its credibility.

Confidence is not corroboration.

Lesson. Inconclusive is not a failure; it is a record waiting for a pattern. The limit: I work in a 600-person organization with a manager who knows employee relations well. A solo HR person may have nobody to ask that question, which is part of why I wanted this written down.

Question for peers. I could not have posted this under my own name: the people involved would recognise it. How do you learn from cases like this when you can’t talk about them?