Build your HR context for AI

Creates a small set of documents about how your company's HR works, so every AI session starts with that context instead of asking you to explain it.

This skill builds a small set of documents that lets any AI session start with your company's HR already understood: how you're organised, when things happen, what's already decided, and where the real documents live. It reads what it can reach, interviews you on the rest, and shows you the plan before it builds anything. It stores patterns and pointers, not employee records.

When to use it

  • You re-explain your company at the start of every AI conversation.
  • Several people or agents use AI for HR and get different answers to the same question.
  • AI keeps reopening decisions you settled long ago, such as who is excluded from review cycles.
  • You're about to use other HR skills and want them to work from one shared source.
  • Policies live in several places and nobody, human or AI, knows which copy is current.

What you get

A small workspace, saved as a folder, as Effy AI knowledge documents, or both:

  • README: a map of what lives where and how to work there.
  • AI usage rules: the rules to read before every task, or a pointer to your policy.
  • company/: org basics, the HR calendar, a decision log, and an index of policies with links to where each one lives.
  • processes/: one document per process you actually run, such as the review cycle, hiring, or onboarding.
  • logs/: what changed in the workspace, and what the AI got wrong and what's true.

Every document stays under 2,000 words. Each inferred line is marked [CONFIRMED] or [GUESS: verify], so you know what to check.

How it works

  1. It picks a home. It checks what's connected and asks once where the context should live: Effy AI knowledge, a shared folder, or both.
  2. It reads what's already reachable. Existing context documents, Effy AI records (counts and structure only), recurring HR events in email and calendar, handbooks and policies in drives and wikis, and the HR, ATS and payroll tools you use.
  3. It interviews you, then shows the plan. Up to ten questions, one at a time, only about what the evidence couldn't answer: who owns each process, which decisions AI should never reopen, which processes cause the most work, and anything that contradicts something else. Then it lists the documents it will create or update and waits for a yes.
  4. It builds and saves. One short document per purpose, saved after an explicit yes, with the first line added to the context log.
  5. It offers two extras. A project pack for each HR job you run, and a weekly review that reads the logs and proposes fixes. The review never edits on its own.

How it weighs evidence. A document or a recurring calendar event counts more than a one-off email. What you said counts more than what it inferred.

What it never does. It doesn't write employee names, salaries, ID numbers, or health details. It doesn't copy policies or templates: it links to the master. It doesn't fill a gap with what companies like yours usually do, and a process you don't run gets no document.

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: connect email, calendar and your document drive, plus a folder to save to. The context can also be stored as knowledge documents in Effy AI.

  1. Ask, for example: "Effy, build our HR context for AI."
  2. Answer the interview questions.
  3. Approve the plan, then approve the save.

For onboarding, the AI usage policy and the handbook, the dedicated skills write a fuller document than this one does; it will offer them where they apply.

Click. Try. Use.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.