Why the question comes up at all

An investigation stands or falls on whether people trust the process. If the outcome looks as though it was decided in advance — or even if there's a reasonable suspicion of that — the findings lose their force. Employees may not accept them, the outcome may be challenged at an employment tribunal, and regulators or clients may take a dim view. Conflicts of interest sit right at the centre of that risk.

The first thing to say, warmly and clearly, is that conflicts are normal. Investigators are drawn from inside the organisation. They have worked with people for years, sat on the same committees, shared managers and occasionally fallen out with each other. The problem isn't the existence of a link; it's a link that nobody noticed, or noticed and quietly ignored. Handled well, a declared conflict is a sign of a healthy process, not a weakness in it.

Where the links tend to hide

Most people think of conflicts as outright corruption or obvious family ties. In practice, they are far more mundane. It's worth briefing your team on the full range, because people genuinely don't always recognise their own.

• Personal: family or household relationships, long-standing friendships, romantic connections past or present, close social ties or shared membership of a tight-knit team.
• Professional: current or recent reporting lines, mentoring relationships, a shared manager, past employment, a promotion or pay decision in which one person influenced the other.
• Financial: shareholdings, bonuses or commission arrangements linked to the subject matter, or a stake in a supplier under scrutiny.
• Historical: a previous grievance or disciplinary process involving the same people, or earlier involvement in the same matter or something closely related to it.
• Attitudinal: a known grudge, rivalry or, conversely, a strong personal loyalty. Bias can push in either direction, so a dislike of the person under investigation deserves just as much attention as friendship with them.

Ask early, ask in writing

Declarations should be collected before anyone reads the papers, not after. Once a team member has absorbed the evidence, it's much harder for them — and for you — to judge whether their view has been coloured.

Use a short, plain declaration form rather than a vague verbal check-in. Ask people to list any relationship with the parties involved, the witnesses, or the subject matter, and to note anything that a fair-minded outsider might raise an eyebrow at. Make clear that declaring something will not be treated as an admission of wrongdoing.

It helps to widen the net beyond the obvious. The lead investigator and the HR adviser are the usual focus, but the same questions apply to note-takers, IT colleagues pulling data, translators, subject matter experts and any external advisers brought in. Ask again at key milestones — before interviews, before findings are drafted, and before any decision on sanction — because circumstances shift.

Keeping a proper record

Every declaration and every decision about it should be written down. A simple conflict register works well, capturing who declared what and when, what was decided, who decided it, and the reasoning behind it.

This record does two jobs. It demonstrates that the organisation took the issue seriously, and it protects the individual investigator, who can point to a documented, considered process if their involvement is later questioned. Keep the register separate from the main investigation file with restricted access, and handle it in line with data protection law — collect only what's relevant, store it securely, and don't retain it longer than necessary.

Deciding what to do about it

Not every declaration requires drastic action. Responses sit on a scale, and the skill is choosing the lightest measure that genuinely protects integrity.

• Log and continue: the link is remote or trivial, and there is no realistic prospect of it affecting judgement.
• Mitigate: keep the person involved but ring-fence their role — no involvement in interviewing particular witnesses, no sight of certain evidence, no part in the decision on findings.
• Reassign: move the specific duty to someone else while the person remains on the wider team.
• Step aside: the person withdraws from the investigation entirely.

Three questions help you choose. Could this reasonably affect the person's judgement? Would a reasonable outsider, knowing the facts, question their neutrality? And is there a workable way to manage it, or is removal simply the cleaner answer? Perception matters as much as reality here, so don't be shy about erring on the side of caution.

Reassigning without derailing things

When someone has to step back, do it well. Identify a replacement early so momentum isn't lost, and plan a careful handover that preserves confidentiality. Avoid broadcasting the reason; a brief, neutral explanation to the wider team is usually enough. And be thoughtful about the person stepping aside — being removed from a high-profile matter can sting, particularly when they did the right thing by declaring. Thank them for it.

Watch, too, for the reverse problem: someone assigned because they have no link at all, but who carries an agenda of their own. A clean conflict check isn't the same as an assurance of objectivity.

Finally, keep the process alive. As evidence emerges, it may implicate someone connected to a team member who declared nothing at the start. Revisit the register, ask again, and document what you decided. A short reminder at each stage, backed by leaders who declare their own interests openly, does more than any policy to make honesty the easy option.