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

If Claude Skills is new to you

Claude Skills are reusable instruction folders that Claude can load when a task matches the skill description. Current Anthropic help says Skills are available on Free, Pro, Max, Team, and Enterprise plans when code execution is enabled.

For this workflow, you will create one private custom skill, upload it in Claude under Customize > Skills, enable it, test that Claude recognizes it, then use it to turn a draft policy change and real staff questions into an internal FAQ. The Claude Skills tool hub is useful further reading, but this page gives you enough setup detail to run the workflow.

Claude is not the policy owner. It can organize questions, rewrite approved language in plain English, and flag gaps. It must not silently resolve ambiguous policy language or invent decisions.

What skill you will use in this workflow

You will use a custom skill called policy-change-faq.

That skill does one narrow job: it turns a draft policy change, manager notes, and real staff questions into a plain-language internal FAQ. It groups repeated questions, rewrites confirmed answers for staff, and keeps unresolved policy points visible for human review.

This task is repetitive enough to deserve a skill because policy changes often follow the same pattern: staff ask what changed, who is affected, when it starts, what they need to do, and what remains unclear. Before using the skill, prepare the current policy draft, known staff questions, audience notes, key facts, and any policy-owner confirmations you already have.

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

• Claude with code execution enabled and access to Skills.
• 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].
• Permission to use the policy material in Claude under your organization's AI and data policy.
• A policy owner who can review unresolved or sensitive points before the FAQ is published.

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.

Build the custom skill for this workflow

Create a folder named policy-change-faq. Inside it, add a file named Skill.md. Paste the starter content below, then adjust it for your organization's house terminology, review markers, and publishing rules.

---
name: policy-change-faq
description: Turn draft policy language, manager notes, and real staff questions into a plain-language internal FAQ while flagging unresolved points for human review.
---

# Policy Change FAQ Builder

Use this skill when the user needs to turn a draft policy change into an internal FAQ for staff or managers.

## What this skill should do

- Group repeated questions into clean FAQ headings.
- Rewrite policy language into plain English without changing the meaning.
- Separate confirmed answers from unresolved policy points.
- Add a review-needed marker when the source material is ambiguous.

## Output shape

- Short intro explaining what changed.
- FAQ entries with one clear question and one clear answer each.
- Final section called 'Needs policy owner confirmation' when unresolved points remain.

## Guardrails

- Do not invent policy decisions.
- Do not claim something is approved unless the source material says so.
- Keep legal or compliance uncertainty visible.
- Prefer plain language over policy jargon.

## Examples

- Turn this draft policy change and these manager questions into an internal FAQ.
- Create a staff FAQ from these policy notes and flag unresolved points.
- Use the policy FAQ workflow and keep unapproved answers out.

Zip the policy-change-faq folder so the ZIP contains the skill folder at the root. Upload the ZIP in Claude under Customize > Skills, then enable the skill. Do not upload a ZIP that contains loose files at the root without the skill folder.

This starter is a first draft, not a final compliance artifact. If your organization uses exact terms, required disclaimers, a review workflow, or a manager-only section, add those rules to Skill.md before you upload it.

Optional official Anthropic skill

Anthropic's official internal-comms skill is a strong fit for FAQ-style internal communication. Use it when you need a general internal update, announcement, or FAQ structure and your source material is already approved.

Use the private policy-change-faq skill when the organization needs house terminology, a mandatory unresolved-items section, fixed approval language, or repeatable staff FAQ rules that should not change from one policy update to the next.

Why this should be a custom skill, not a one-off prompt

Policy FAQs are high-risk because the same pattern repeats but the facts change. A custom skill keeps the fixed structure, review markers, plain-language rules, and no-invention guardrails in one reusable place. That lowers prompt rewrite friction and makes it harder to forget the unresolved-items section.

Skill trigger test

After enabling the skill, try these prompts with a small policy sample:

• Create an internal FAQ from this draft policy update.
• Group these manager questions into a staff-facing FAQ.
• Use my policy FAQ skill and flag anything that needs policy owner confirmation.
• Turn these policy notes into a plain-language FAQ without inventing answers.

Claude should treat these as policy-change-faq tasks and include a clear unresolved-items section when the source is incomplete. If it does not, tighten the skill description so it names internal FAQs, draft policy updates, staff questions, and policy-owner confirmation.

Numbered workflow steps

1) Gather and clean the source material

Collect the current draft policy, manager notes, staff questions, go-live date, affected audience, and known exceptions. Remove duplicates and conflicting old drafts unless you need the policy owner to compare them.

Create or update:

• faq-audience.txt for staff group, reading level, and tone
• key-facts.txt for go-live date, owner, affected roles, and required actions
• staff-questions.txt with one question per line
• confirmed-answers.txt for answers the policy owner already approved

2) Create, upload, and enable the custom skill

If you have not already done it, create the policy-change-faq folder, add Skill.md, zip the folder, upload it in Claude under Customize > Skills, and enable it.

If the skill already exists, open Claude's Skills settings and confirm it is enabled before starting the policy FAQ run.

3) Run a quick trigger test

Use one short prompt before uploading the full policy package:

Use my policy FAQ workflow to turn this into a staff FAQ:

Policy change: expense reports now require receipts for all meals.
Staff questions: What if I lost a receipt? Does this apply to travel already booked?
Known fact: effective date is May 1.

Check that Claude groups the questions, answers only from the stated facts, and marks the lost-receipt and already-booked-travel issues as needing policy owner confirmation.

4) Run the actual workflow with the skill active

Upload or paste the policy draft, faq-audience.txt, key-facts.txt, staff-questions.txt, and any confirmed answers. Then ask:

Use the policy-change-faq skill to turn this draft policy change and staff question list into an internal FAQ.

Use only the attached or pasted source material. Write in plain internal language. Keep unresolved policy points visible as [POLICY OWNER TO CONFIRM]. Do not invent exceptions, approval paths, dates, or legal interpretations.

If Claude asks for missing policy details, answer only with approved information. If the answer is not approved, tell Claude to keep the item in the unresolved section.

5) Review unresolved points with the policy owner

Before publishing, send the policy owner the FAQ draft and the Needs policy owner confirmation section. Confirm:

• go-live date
• affected audience
• exceptions
• approval paths
• manager-only guidance
• any legal, HR, benefits, or compliance wording

Claude should not decide these points. It should preserve them for review.

6) Publish, announce, or reuse the result

After policy-owner review, update the FAQ with confirmed answers. Then publish it in the approved internal channel and save the final FAQ, source files, and any useful corrections in the same team folder or Claude Project.

If the review exposed a recurring rule, update Skill.md before the next policy change so the skill gets better.

Tool-specific instructions

Claude Skills

• Use the page-specific policy-change-faq custom skill first.
• Keep internal-comms in mind as the closest official public skill for broader internal communication and FAQ-style drafting, not as a full replacement for your private policy FAQ rules.
• Anthropic's official skill-creator skill can help draft, test, or refine your custom skill if you want help improving the description, trigger prompts, or guardrails.
• For broader setup context, examples, and related Claude Skills workflows, see the Claude Skills tool hub.

Low-friction fallback

If you cannot upload a skill yet, keep the starter Skill.md text, one approved FAQ example, faq-audience.txt, and key-facts.txt in a Claude Project. Paste the skill instructions before each run. This is less reliable than a real skill because you must remember the setup every time.

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.
5. Nothing presents Claude as the policy owner or final approver.

Common failure modes and fixes

The FAQ sounds too formal

Fix: update Skill.md to prefer plain internal language over policy jargon. Add one approved example answer if your organization has a house style.

Claude hides unresolved items

Fix: stop the draft and rerun the trigger test. Update the skill to require a final Needs policy owner confirmation section whenever the source material is incomplete or ambiguous.

The FAQ grows too long

Fix: add a length rule to the skill. For example: "Keep each staff answer to 2 to 4 sentences unless the user asks for manager detail."

Claude invents an exception or approval path

Fix: remove the invented answer before publishing. Add the issue to the policy-owner review list and update the skill guardrails to keep unapproved exceptions marked [POLICY OWNER TO CONFIRM].