How to Use Claude Skills to Turn a Draft Policy Change Into an Internal FAQ

Warning: If the policy update affects HR, legal, benefits, health information, or disciplinary procedures, only use Claude in a workflow your organization has approved.

If Claude Skills is new to you

Start with the Claude Skills explainer page in the site's Tools menu before you run this workflow. The short version is that Skills are reusable instruction bundles Claude can load for repeatable work. The public skills repository also makes good reference material when you want to copy the pattern into a private skill for your own team. For curated examples, see the Claude Skills tool hub and the Anthropic Workflows series.

Problem and who this is for

A policy owner gives you a draft update. Staff immediately start asking the same practical questions: what changed, when it starts, who is affected, what people need to do, and what stays the same. A policy document answers some of that, but not in the form staff actually want.

This workflow is for office admins, school and clinic administrators, executive assistants, HR support staff, and coordinators who need to turn a policy change into an internal FAQ that people can actually use.

Prerequisites

• A paid Claude plan if you want to use the example Skills path in Claude.ai.
• The latest draft policy in a readable format such as DOCX, PDF, or pasted text.
• A rough list of real staff questions from email, chat, meeting notes, or your own experience.
• A simple answer governance rule. Example: unresolved items stay marked [POLICY OWNER TO CONFIRM].

How to gather the source material

1. Export the policy draft into one clean version. Do not feed Claude three conflicting drafts if you already know which one is current.
2. Pull staff questions into a single list. Use plain text and one question per line.
3. Remove duplicates and near-duplicates. Keep the plainest wording employees would really use.
4. Add one short note called faq-audience.txt that says who the FAQ is for and what tone it should use.
5. If the policy has a go-live date, owner, or exceptions, put those in a small key-facts.txt file so Claude does not miss them.

Numbered workflow steps

1) Use a Skills workflow built for internal communications and structured drafting

This article works best with a Claude Skills workflow inspired by both internal-comms and doc-coauthoring. The first gives Claude a strong pattern for internal FAQs. The second is useful when the policy is incomplete and the wording needs iteration before publishing.

2) Make Claude classify the questions before answering them

{
  "task": "Classify staff questions for a policy-change FAQ",
  "input": {
    "policy_draft": "ATTACHED FILE OR PASTED TEXT",
    "staff_questions": "PASTE QUESTION LIST",
    "faq_audience": "PASTE faq-audience.txt",
    "key_facts": "PASTE key-facts.txt"
  },
  "rules": [
    "Do not answer yet.",
    "Group duplicates under one normalized question.",
    "Identify unanswered questions that the draft policy does not clearly resolve.",
    "Mark anything ambiguous as [POLICY OWNER TO CONFIRM]."
  ],
  "output": {
    "grouped_questions": "Question groups with duplicate variations listed under each",
    "clear_answers_available": "Questions the policy clearly answers",
    "needs_confirmation": "Questions that require policy owner review",
    "missing_topics": "Important practical questions staff are likely to ask that are not yet on the list"
  }
}

3) Review the gaps before Claude writes the FAQ

This is the step people skip and regret later. If something is ambiguous, confirm it now. A good FAQ is not just well written. It is operationally safe.

4) Generate the first FAQ draft

{
  "task": "Write an internal FAQ from a policy change and a verified question set",
  "input": {
    "verified_question_groups": "PASTE THE REVIEWED OUTPUT FROM STEP 2",
    "policy_draft": "ATTACHED FILE OR PASTED TEXT",
    "audience": "STAFF AUDIENCE HERE",
    "tone": "plain language, direct, internal"
  },
  "rules": [
    "Write in plain internal language, not legal language.",
    "Lead with what changed, when it starts, and who is affected.",
    "Use one short answer per question unless more detail is truly needed.",
    "Keep unresolved items visible with [POLICY OWNER TO CONFIRM].",
    "Do not invent exceptions or approval paths."
  ],
  "output_format": {
    "type": "faq",
    "sections": [
      "What changed",
      "Who this affects",
      "FAQ",
      "What staff should do next",
      "Questions still awaiting confirmation"
    ]
  }
}

5) Create a shorter announcement version

Once the full FAQ is good, ask Claude for a short announcement paragraph that links people to the FAQ. That gives you both the reference document and the launch message in one pass.

Tool-specific instructions

Claude Skills

• If Skills is new to you, read the Claude Skills explainer page in the site's Tools menu first.
• Claude can automatically load relevant skills when the task matches. That is useful here because the work looks like an internal FAQ and a structured document-drafting task.
• If you want a private version, Claude's help documents custom skill upload by ZIP file.
• The public skills repository is useful for pattern borrowing even if you do not upload anything right away.

Quality checks

1. Every FAQ answer can be traced back to the policy draft or your manual confirmation notes.
2. The go-live date, affected audience, and required action are identical across the announcement and FAQ versions.
3. Any unresolved point remains marked for confirmation instead of being smoothed over.
4. The answers sound like staff guidance, not like policy prose copied without translation.

Common failure modes and fixes

The FAQ sounds too formal

Fix: rerun Step 4 with the tone set to plain internal language and ask Claude to write like an experienced operations admin, not a policy writer.

Claude hides unresolved items

Fix: restate the rule that unresolved items must remain visible with [POLICY OWNER TO CONFIRM].

The FAQ grows too long

Fix: split the output into an employee FAQ and a manager-only FAQ, or ask Claude to keep each answer to 2 to 4 sentences.

Sources Checked

• Anthropic Claude Help Center: What are Skills? (accessed 2026-03-23) https://support.claude.com/en/articles/12512176-what-are-skills
• Anthropic Claude Help Center: Use Skills in Claude (accessed 2026-03-23) https://support.claude.com/en/articles/12512180-use-skills-in-claude
• Anthropic Claude Help Center: How to create custom Skills (accessed 2026-03-23) https://support.claude.com/en/articles/12512198-how-to-create-custom-skills
• GitHub: elliepetalmedia/skills README (accessed 2026-03-23) https://github.com/elliepetalmedia/skills
• GitHub: internal-comms skill in elliepetalmedia/skills (accessed 2026-03-23) https://github.com/elliepetalmedia/skills/tree/main/skills/internal-comms
• GitHub: doc-coauthoring skill in elliepetalmedia/skills (accessed 2026-03-23) https://github.com/elliepetalmedia/skills/tree/main/skills/doc-coauthoring

Quarterly Refresh Flag

Review on 2026-06-21 to confirm any updates to Claude Skills, Cowork availability, mobile dispatch requirements, project behavior, and supported setup paths.