Day 7: Write a Policy Document From a Brief

The concept

Policies are not just documents. They are operating instructions for moments when people need clarity.

A good HR policy helps employees understand what applies to them, helps managers respond consistently, and helps the organisation show that decisions are made against a clear standard. A poor policy does the opposite. It hides simple rules inside dense language, leaves managers guessing, and creates extra work for HR every time a normal situation becomes confusing.

AI can help draft policy documents from a brief. But this is an area where HR judgement matters heavily. The model can organise, simplify, and suggest language. It cannot decide legal entitlement, contractual obligation, local compliance, or organisational risk appetite.

Plain English

The brief is the thinking. The AI draft is the structure. Legal review is the safety check.

Start with the brief

Most policy work gets stuck because the organisation has not made the decisions the policy needs to express.

Before asking AI to draft anything, write a short brief.

Your brief should answer:

• Who does the policy apply to?
• What is the rule, entitlement, or process?
• What must the employee do?
• What must the manager do?
• Who approves exceptions?
• What records are kept?
• What happens if the policy is misused?
• What legal or contractual constraints apply?
• When should the policy be reviewed?

If those answers are missing, AI will fill gaps with plausible language. That can be dangerous. A polished policy built on invented assumptions is worse than a rough draft that shows where decisions are needed.

Make the document usable

Policy writing often becomes legalistic because people confuse seriousness with complexity.

A policy can be legally important and still be written clearly. In fact, clarity usually improves the policy's usefulness.

Employees should be able to find:

• whether the policy applies to them
• what they are allowed or required to do
• how to make a request
• who approves it
• what happens next
• where to go with questions

Managers should be able to find:

• their responsibilities
• decision points
• escalation triggers
• documentation requirements
• how to handle exceptions

If people cannot use the policy under pressure, the document has failed.

Use a consistent structure

The prompt asks for a defined structure because consistency matters.

Purpose

Why does this policy exist?

Scope

Who is covered and who is not?

Policy Statement

What is the organisation's position?

Eligibility

Who qualifies, and under what conditions?

Process

What steps should the employee, manager, and HR follow?

Manager Responsibilities

What must managers do consistently?

Employee Responsibilities

What must employees do?

Records

What should be documented and where?

Exceptions

Who can approve exceptions, and how should they be handled?

Review Date

When will the policy be reviewed?

This structure turns policy into a usable reference rather than a long narrative.

Plain English is a control, not a compromise

Plain language reduces misinterpretation.

For example, instead of:

Employees are required to submit requests in accordance with the procedural provisions outlined herein.

Write:

Employees must submit requests using the process in this policy.

The second version is not less professional. It is clearer.

Plain language is especially important for policies that employees read during stressful moments: sickness, leave, grievances, flexible working, performance, disciplinary matters, redundancy, or complaints.

When people are anxious, clarity matters more.

Know what AI should not decide

AI should not be treated as legal advice.

Human review is needed for sections involving:

• statutory entitlements
• protected characteristics
• contractual obligations
• disciplinary consequences
• dismissal risk
• pay or benefits
• leave rights
• data privacy
• health and safety
• jurisdiction-specific requirements

The model can identify where review is needed, but your organisation must decide who signs off.

If the policy affects employment rights, legal or compliance review is not optional.

Add a manager FAQ

Policies often fail at the manager layer. HR publishes a policy, managers receive questions, and inconsistent answers appear almost immediately.

A manager FAQ helps prevent that.

Good FAQ questions include:

• What if an employee asks for an exception?
• What if the request affects team coverage?
• What if the employee disagrees with the decision?
• What records should I keep?
• When should I involve HR?

The FAQ should be practical, not theoretical. Managers do not need a legal lecture. They need a safe path for common scenarios.

Plan the rollout

A policy does not become real because it is uploaded to the handbook.

For a useful rollout, decide:

• who needs to know first
• what managers need before employees ask questions
• whether the policy replaces old practice
• what message employees should receive
• where the policy will live
• how questions will be handled
• when feedback will be reviewed

AI can draft the rollout note, but HR should make sure the rollout matches the organisation's culture and communication channels.

Test the policy with a scenario

Before publishing, test the policy against two or three realistic scenarios.

Ask:

• Can an employee understand what to do?
• Can a manager apply the process consistently?
• Does HR know when to intervene?
• Are exceptions handled clearly?
• Are records and privacy addressed?
• Does any wording create a promise we do not intend to make?

If a scenario breaks the policy, fix the policy before rollout.

Create a review rhythm

A policy should not be treated as finished forever. Laws change, working patterns change, technology changes, and informal practice often drifts away from the written document.

Set a review rhythm when the policy is created. For many HR policies, an annual review is sensible. For policies affected by fast-moving regulation, employee relations risk, data privacy, or major organisational change, review more often.

At each review, ask:

• Is the policy still legally current?
• Does the written process match what managers actually do?
• Are employees asking the same clarification questions repeatedly?
• Have exceptions become common enough to suggest the policy is unrealistic?
• Are records being kept in the right place?
• Has any recent case exposed a gap in wording or process?

This governance habit protects the organisation from stale documents and protects employees from inconsistent practice.

AI can help create the review checklist, summarise feedback, and propose clearer wording. But the review decision should remain with HR, legal, and accountable business leaders.

Today's practice

Choose one policy that is missing, outdated, or frequently misunderstood. Write the brief before running the prompt.

Then review the output:

1. Where did AI make an assumption?
2. Which section needs legal or compliance review?
3. Would an employee understand the process in two minutes?
4. Would a manager know when to involve HR?
5. What scenario should we test before publishing?

By the end, you should have a policy draft that is clearer, more usable, and safer to take into review.