Something happened at work that concerns you — a comment that crossed a line, a manager who keeps changing your schedule without notice, a safety shortcut you have been asked to take repeatedly. You are not sure yet whether it needs a formal complaint, but you know from experience that memory fades and details blur, so you want to write it down properly while it is fresh. AI is a useful tool for this specific, narrow task: turning your memory into a clean, organized, dated record. It is not a tool for answering the question underneath — is this actually against the rules, and what should I do about it? — because that question requires knowledge of your specific employer’s policy, your jurisdiction’s law, and facts a general-purpose model has no way to verify.
This is a literacy article, not legal advice, and it stays that way deliberately. Nothing here tells you whether a specific situation is illegal, what any specific law entitles you to, or how a complaint should be worded to succeed. That is the entire point: those questions belong with people who are actually positioned to answer them for your specific situation.
A private note does not pause reporting, grievance, appeal, limitation, evidence-preservation, or safety deadlines. For imminent danger, criminal conduct, safeguarding concerns, or a deadline, use the appropriate qualified channel immediately rather than continuing to document with AI.
Why a factual record matters, regardless of what happens next
A prompt factual note can preserve details that may otherwise be forgotten: exact dates, direct quotes as remembered, people present, and references to existing messages or documents. It is still one person’s account, not automatically proof, and its evidential treatment depends on the applicable process and law. Preserve original documents and metadata rather than replacing them with an AI-organized summary.
Never ask an AI tool whether a specific situation is legally harassment, discrimination, retaliation, or a violation of a specific law, and never ask it to draft a formal legal complaint or tell you what you are entitled to. These require knowledge of your employer’s actual policies, your specific jurisdiction’s law, and case-specific judgment that a general-purpose model cannot reliably provide — and an incorrect answer here can cost you more than no answer at all. Take these questions to your HR department, your union representative, or a licensed employment lawyer in your jurisdiction.
What AI can actually help with
The narrow, useful task is organizing your own account into a clear structure, not generating or interpreting content:
I am going to describe something that happened at work. Organize
what I tell you into a clean, chronological, factual log with these
fields for each entry: date, time, location, who was present, what
was said or done (quote directly where I give you a quote), and any
documents or messages related to it. Do not add interpretation,
characterize the events as any category of misconduct, or suggest
what I should do about it — just organize the facts I give you into
this structure.
Here is what happened: [your account]
Read the output back against your own memory. Correct anything that drifted from what you actually said, and remove anything that reads as interpretation rather than fact — a model can slip a word like “aggressively” or “clearly retaliatory” into a summary even when your original account was more neutral, and that kind of characterization does not belong in a factual log.
A workplace issue log often contains sensitive personal information about you and about other people — health details, personal circumstances, allegations about someone’s conduct. Treat this document with the same caution as work secrets: if your employer has an approved, enterprise-grade AI tool with strong data-handling terms, use that; if not, consider drafting this document without AI at all, or using a local tool that does not send your text to a third party, given how sensitive the content usually is.
What to log, specifically
| Field | What to include |
|---|---|
| Date and time | As precise as you can recall; note if approximate |
| Location | Where it happened |
| Who was present | Names, roles — note if anyone else witnessed it |
| What was said or done | Direct quotes where possible, otherwise as close to verbatim as you can recall |
| Your response | What you said or did at the time |
| Related documents | Emails, messages, schedules, photos — reference where they are saved |
| Pattern, if any | Whether this is a first occurrence or part of a repeated pattern, with dates of prior instances |
Keep opinions, legal conclusions, and characterizations (“this was harassment,” “this was clearly retaliation”) out of the log itself. State what happened; let the people responsible for evaluating it — HR, a union representative, or a lawyer — apply the relevant standard.
Where this goes next is not an AI question
Once you have a clean factual record, the decision about what to do with it — raise it informally, file a formal complaint, consult a lawyer, or wait and keep documenting — is a judgment call that depends on your specific employer’s process, your relationship with the people involved, and factors only you can weigh. A few starting points, offered without assuming your specific situation:
- Your employer’s HR department or designated process may be one formal channel — check your employee handbook or intranet for the specific procedure, the same way finding your employer’s AI policy covers locating the actual document rather than guessing. It is not automatically independent or the safest first contact in every case, especially when HR, senior management, retaliation risk, criminal conduct, or safeguarding is part of the concern; a union, regulator, safeguarding lead, emergency service, or lawyer may be the appropriate route.
- A union representative, if you are represented, can explain the collective agreement and process and may be able to accompany you. In the United States, the NLRB publishes a current Weingarten rights overview; do not transfer that US rule to another jurisdiction or assume it applies to your facts.
- A licensed employment lawyer in your jurisdiction is the right source for any question about what a law actually requires or permits in your specific case — a factual log like the one above is exactly the kind of material a lawyer would want to see first.
For a low-risk issue with no deadline, documenting while you consider options may be reasonable. That sentence does not apply where safety, safeguarding, mandatory reporting, retaliation risk, or a filing deadline requires prompt qualified action.
Two things people get wrong about documenting
The first is treating a well-organized AI-assisted summary as equivalent to legal or HR guidance, because it reads clearly and confidently. A model producing a tidy timeline has not evaluated your situation against any actual policy or law — it has organized text. The clarity of the output says nothing about whether the underlying situation meets any legal or policy threshold, and treating fluent organization as expert judgment is exactly the gap this article exists to close.
The second runs the other way: assuming documentation is necessarily a formal accusation. A factual note can be private preparation, but it does not guarantee protection or a legal outcome and must not replace a required report.
Jurisdictional starting points include the EU Commission’s employment-equality overview, Estonia’s Labour Inspectorate and Equal Treatment Commissioner, and the US EEOC employee portal. Use the applicable employer, union, regulator, safeguarding, or legal process; do not combine these regimes.
Start the record now, while it’s fresh
Use the workplace issue facts log to capture what happened while your memory is clearest, using AI only to organize the structure, never to characterize or advise. When you are ready to decide what to do with it, take it to HR, your union representative, or a licensed employment lawyer — not back to the AI tool that helped you write it down.
