P
Free tool · No signup · Generated in your browser

Company AI Usage Policy Generator

A first-draft, ten-article company AI usage policy matched to your industry, approved tools, sensitive data (GDPR) and the EU AI Act. Copy it or download it as Markdown (.md), plain text (.txt) or PDF.

Configure your policy

Live preview · Generated in your browser

The name stays in your browser: it is never sent or stored.
Team size
Approved commercial AI tools
Sensitive data to restrict (Red Zone)
Human review requirement
Person to notify in case of an incident or new tool request.

Your policy draft (10 articles)

Ready to paste into Notion, Word or Google Docs

# Artificial Intelligence Usage Policy — [Company Name]

> **Working Draft** — Sector: Services and general SMB operations. Review and adapt with legal counsel prior to formal adoption.

## Article 1 — Purpose and Scope
This policy defines the rules for using generative AI tools (chat assistants, copilots and automation workflows) within **[Company Name]**. It applies to all employees, executives, interns and contractors acting on behalf of the organization.

## Article 2 — Core Principles
- **AI assists, humans decide**: AI tools never replace professional judgment or individual accountability.
- **Accuracy and honesty**: never present unverified AI claims or fabricated citations as established facts.
- **Data minimization**: provide only the minimum context required to complete the task.

## Article 3 — Approved Tools and Shadow AI Prevention
Only the following company-approved environments may be used for business tasks:
- ChatGPT Business / Enterprise (company workspace, no training on data by default)
- Claude Team / Enterprise (Anthropic commercial workspace, conversation ratings turned off by the admin)

- **Personal accounts prohibited for internal data**: free or personal consumer accounts (which may retain chats for model training) must never receive company data.
- **New AI tools** (browser extensions, third-party SaaS, plugins or connectors) require prior approval from management or the designated data protection contact.

## Article 4 — Data Classification (Green / Amber / Red)
- **Green Zone (permitted)**: public information, generic outlines, anonymized brainstorming and neutral copy edits.
- **Amber Zone (permitted only on approved commercial tools after anonymization)**: internal meeting summaries, drafts and document reviews stripped of direct identifiers (replace names, clients and exact figures with `[CLIENT_A]`, `[AMOUNT]`).
- **Red Zone (strictly prohibited in external prompts)**:
- Identifiable client records, contracts, negotiated pricing and confidential correspondence
- Resumes, employee files, performance reviews, compensation and hiring data
- Non-public financial statements, tax filings, bank details and forecasts
- **Credentials, API keys and access tokens**: strictly prohibited in all AI tools.

## Article 5 — Permitted and Prohibited Uses
- **Encouraged uses**: structuring ideas, drafting outlines, rewriting copy, anonymized meeting notes, internal documentation and research preparation.
- **Prohibited uses**: fully automated decisions with legal or significant personal impact without human review; generating fake testimonials or records; bypassing internal security controls.
- **Hiring and HR decisions (EU AI Act, Annex III)**: automated candidate filtering, applicant scoring, promotion or termination decisions must never be delegated to an AI tool. AI may only assist with drafting structured templates reviewed by humans. Annex III high-risk obligations apply from December 2, 2027 (Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744).
- **Figures, calculations and legal citations**: every calculation, amount, statutory reference or deadline produced by an AI tool must be independently recalculated and verified against primary official sources.

## Article 6 — Human Review and Intellectual Property
No AI-generated output may be sent to a client, partner or candidate without thorough human review and ownership by an identified team member.
- **Copyright and trademarks**: do not paste copyrighted third-party material without permission and verify deliverables before publication.

## Article 7 — Security, Privacy and Incident Reporting
- **Account settings**: keep model training opt-outs active and avoid submitting feedback ratings (“thumbs up/down”) on internal conversations.
- **Accidental disclosure**: if Red Zone data is pasted into an unapproved tool by mistake, notify management or the designated data protection contact immediately so the chat can be deleted and any required GDPR breach assessment (72-hour window) can be conducted promptly.

## Article 8 — AI Literacy and Training (EU AI Act, Article 4)
In line with Article 4 of the EU AI Act (Regulation (EU) 2024/1689, applicable since February 2, 2025), **[Company Name]** supports role-appropriate AI literacy and practical training covering prompting methods, model limitations, confidentiality and output verification.

## Article 9 — Governance and Updates
Under GDPR Article 37, where no mandatory DPO is appointed, oversight is handled by the designated data protection contact or external legal counsel. This policy is reviewed at least once a year or whenever tools or regulations change materially. Internal contact: **management or the designated data protection contact**.

## Article 10 — Acknowledgment
Each team member acknowledges reading this policy and commits to following its confidentiality, anonymization and human-review rules when using AI tools at **[Company Name]**.

---
_Generated for free on prompt-guide.com/en/entreprises/generateur-charte-ia — Review with legal counsel before formal adoption._

The Green / Amber / Red rule

A long policy ends up in a shared folder. What staff remember at their desk is a three-color rule for sorting data.

Green Zone · Permitted

Public or generic material

Brainstorming, rewriting public copy, outlines, translating non-confidential text and generic templates.

Amber Zone · Anonymized + business tier

Sanitized internal drafts

Meeting notes or sales drafts, only inside an approved business workspace and after replacing client names and exact figures.

Red Zone · Strictly prohibited

Sensitive data & secrets

Passwords, API keys, identifiable resumes and HR records, health data, tax filings, NDA-covered contracts or critical source code in unapproved external tools.

4 checks before formally adopting your policy

  • Have legal or HR counsel review disciplinary, IP and works-council provisions prior to adoption.
  • Run the GDPR checklist from our security comparison before opening accounts.
  • Name an internal contact people can ask for a new tool or tell about an accidental paste.
  • Walk the team through the policy for an hour or two on real cases instead of sending a PDF.
Quentin Tousart

Team training & rollout support

Train your team on your real deliverables

Quentin Tousart (founder of Prompt Guide and CEO of Webpulser) helps SMBs, consulting practices and training organizations put AI to work. Quoted per project, based on team size and goals.

Hands-on team workshop (half-day)

Practical session built around your real tasks, prompting frameworks and confidentiality rules.

Custom in-house training day

Role-specific tracks (HR, sales, finance, ops), internal prompt library and usage policy.

Workflow & rollout advisory

Structuring recurring AI workflows on Claude, ChatGPT or Gemini with your teams.

Book a 15-min intro callAbout the founder

Looking for self-serve role workflows instead? Explore our ready-to-use business kits on AI Cowork Kit.

AI usage policy FAQ

Is a written AI usage policy legally mandatory for SMBs?

No law requires a document called an “AI policy”. Two EU texts still push you to write one: GDPR, for personal data shared with an AI processor, and EU AI Act Article 4, applicable since February 2, 2025, which asks organizations to act on staff AI literacy.

Is any data entered into this generator stored?

The policy is assembled in your browser and is not stored anywhere. Your company name, contact person and the policy text never leave your device. If you accepted audience measurement, we send Google Analytics an event when you copy or download the policy, with the sector and team size you picked, the format and the language. Without that consent, the tool sends no event.

Does this template replace legal review?

No. It is a working draft. Have legal counsel or your HR or data protection lead review it before you adopt it or add it to an employee handbook.

How do you make sure employees actually follow the policy?

Give everyone a business account, or the free personal one stays the habit. Add a one-page Green/Amber/Red summary, then spend an hour or two anonymizing real examples together.

Continue in the Business AI guide