Creates a one-page policy of plain-language rules for how your HR team may use AI, with the points that need legal sign-off marked.
Write your HR team's AI usage policy
This skill writes a one-page AI usage policy for an HR team: numbered rules in plain language, each with a one-line reason. It aligns the rules with your existing company policies and with where your employees and candidates sit, and marks every line that needs legal, DPO or leadership sign-off. It is written to encourage daily use of AI, not only to restrict it.
When to use it
- Your HR team already uses AI and nobody has written down where the lines are.
- People avoid AI because they aren't sure what's allowed.
- Someone pasted employee data into a tool and you want rules before it happens again.
- Legal or security asked for a policy and you want a draft they can review, not a blank page.
- You employ people in both the US and the EU and need one standard.
What you get
One page, with:
- A header: version, effective date, owner, and a review date six months out.
- What we expect: use AI for routine work, share what works, and remember that whoever ran the tool owns the output.
- Numbered rules, covering at least: approved paid accounts with model training turned off; names out by default; a "never enter" data list; what may be entered de-identified; AI never makes employment decisions; approval and bias checks for anything touching candidates or employees; comp numbers from your own data, never from AI; human review for anything legal or going into an employee's file; policy answers that cite the policy; how AI use is disclosed; where deliverables are saved.
- An incident section: what to do the same day sensitive data goes into a tool by mistake, with no blame.
- A sign-off table: legal, DPO, HR lead, leadership, as applicable.
Lines that need sign-off are marked [REVIEW].
How it works
- It asks one question first. Where do your employees and candidates sit: US only, EU/EEA only, both, or somewhere else. Everything else depends on the answer.
- It collects existing policies. It searches connected drives, wikis and Effy AI knowledge for your handbook, data protection, information security, confidentiality, acceptable use and retention policies, and asks you to upload what's missing.
- It interviews you. Up to nine questions, one at a time: who uses AI and for what, which accounts and tools are approved, what data may never be entered, when real names are allowed, what needs human review, how you disclose AI use, and what happens when something goes wrong. It skips anything your existing policies already answer.
- It writes the policy. Where a rule extends an existing policy, the reason names it. Where two policies would conflict, it writes the stricter rule, names the conflict, and marks it
[REVIEW]. For US and EU teams together, it applies the stricter rule everywhere. - It revises until you approve, then saves. Where your team keeps HR documents, and in Effy AI knowledge if you say yes.
What it never does. It isn't a lawyer and says so once. Laws in this area change often, so every legal reference is checked for current status and marked [REVIEW]; the policy needs sign-off from qualified people before it takes effect. It doesn't silently pick a side when policies conflict.
How to run it
Install the Effy AI plugin for Claude or ChatGPT. One install adds every Effy AI skill, this one included. Then connect the tools you want the skill to read from, such as email, calendar, your drive or your meeting notetaker, and ask for what you need in plain words: the assistant picks the right skill.
Before you start: have your existing policies to hand, or connect the drive or wiki where they live.
- Ask, for example: "Effy, write our HR team's AI usage policy."
- Say where your employees and candidates sit, then answer the interview.
- Review the draft, approve it, and send the
[REVIEW]items to the people who sign them off.