Home · Business · HR & People · Policies & handbooks

YES

As of 13 August 2026, AI can create a code of conduct policy for your UK business.

This still needs a person who signs their name to it.

Can you do it?

5 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ita colleague

What the alternative costsNo priced alternative is supplied in the available tool data.

If this goes wrong, staff may receive inconsistent treatment or be unable to report misconduct through a workable route.

What to actually do

  1. Hand it to a person

    The route this page recommends

    A person who owns the outcome does this end to end, worth it when the failure is dear.

  2. Use a tool built for this

    Second choice
  3. Do it yourself

    The distant third

    A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

    How to actually do it

    1. Open your current handbook, disciplinary and grievance procedures, whistleblowing arrangements, equality commitments and health and safety rules, then gather the sections that the new policy must match.
    2. Write down the business name, workforce types, workplaces, working arrangements, reporting contacts, investigation route, disciplinary route and policy owner.
    3. Paste those details into the prompt, replacing each bracketed slot and removing any information the policy should not contain.
    4. Ask the chatbot to produce the policy, implementation checklist and unanswered business decisions in separate sections.
    5. Compare every reporting route, contact, process and consequence in the draft with your current documents, and replace each [BUSINESS DECISION REQUIRED] item before circulation.
    6. Check the draft against current GOV.UK and other official UK employment guidance, then ask an employment solicitor to review provisions where a mistake could affect disciplinary action, discrimination, harassment or dismissal.
    7. Send the approved version to the relevant workers, record how it was communicated, and store the review date and policy owner with the final document.

    Prompt

    Create a UK workplace code of conduct policy for [BUSINESS NAME], a [BUSINESS TYPE] with approximately [NUMBER OR RANGE] workers in [LOCATION].
    
    Use these business details and existing rules:
    [PASTE BUSINESS VALUES, EXISTING POLICIES, REPORTING ROUTES, DISCIPLINARY PROCESS, WORKING ARRANGEMENTS AND ANY COLLECTIVE AGREEMENTS]
    
    The policy should cover:
    - expected standards of behaviour, including respect, inclusion, honesty, confidentiality and conflicts of interest
    - unacceptable conduct, including bullying, harassment, discrimination, retaliation, fraud, misuse of company property and unsafe conduct
    - how workers should raise a concern, including confidential options where available
    - how the business will respond, investigate and protect people from retaliation
    - possible consequences, while making clear that each case depends on its facts and the applicable employment process
    - who the policy applies to, where it applies and how it will be communicated and reviewed
    
    Write in plain British English for employees and workers. Do not invent legal duties, reporting channels, contacts, penalties, benefits or business procedures. Mark every missing decision as [BUSINESS DECISION REQUIRED]. Distinguish clearly between a code of conduct and a disciplinary or grievance procedure. Identify provisions that should be checked against current UK employment law and official guidance. Include a short implementation checklist and a list of questions I must answer before publishing. Do not present this as legal advice.

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

What it gets wrong

  • AI cannot decide which standards fit your workplace, risk profile and working relationships.
  • AI cannot know whether the reporting and investigation routes in the draft actually work in your business.
  • AI cannot reliably resolve every interaction between the policy, employment law and your existing procedures.
  • AI does not carry responsibility if the policy is discriminatory, unclear or applied inconsistently.
  • This is not professional advice. A serious employment-law risk needs review by an employment solicitor.

Even on a YES, the friction has a name: legal accountability, judgement under ambiguity and stakes of error.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability1
Effort delta2
Total8 / 10

FAQ

Can ChatGPT write a code of conduct policy?
Yes, it can produce a useful first draft with sections for expected behaviour, unacceptable conduct, reporting and implementation. You must supply your real procedures and check every contact, process and consequence before publishing.
Is an AI-written code of conduct policy legal in the UK?
AI authorship does not make a policy compliant or non-compliant. This is not professional advice, and an employment solicitor should check provisions that could affect discrimination, harassment, disciplinary action or dismissal.
What should a UK code of conduct include?
It should set out the behaviour your business expects, examples of unacceptable conduct, how concerns can be raised, how they will be handled and how the policy relates to other workplace procedures. The details must match your actual reporting and disciplinary arrangements.
Can I use an AI policy without HR approval?
You can use AI to prepare a draft, but publishing it without checking your procedures creates avoidable risk. Have a responsible manager or HR colleague verify the operational details, and obtain employment-law advice where the consequences are serious.

Nearby answers

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.