AI Policy for Your Company: Template and What It Must Include

Smernica o používaní AI vo firme: vzor a čo v nej musí byť
An AI policy for your company is a short internal document (2 to 4 pages is enough) that tells employees which AI tools they may use, what data they may enter into them, and who is responsible for what. For a company of 5 to 50 people, it is a practical way to introduce AI into your business without the risk of personal data leaks and with proof that you have rules in place.

Why your company needs one

Every company whose employees use ChatGPT, Gemini, Copilot or any other AI tool needs a policy. There are two reasons: the AI Act and the GDPR.

AI Act, Article 4 (AI literacy). Since 2 February 2025, companies that use AI must take measures to promote the AI competence of their employees. A certificate is not required, but during an inspection, evidence is expected, for example a policy and a training record. The Digital Omnibus (Regulation (EU) 2026/1744, in force since 27 July 2026) did not repeal this article, it only reworded it into a more flexible form. Details are in Blck Alpaca’s analysis.

GDPR. Every entry of personal data into an AI tool is processing of personal data. For this you need a legal basis, a contract with the tool’s provider, and a clear rule on what must not be entered into AI.

The current AI Act deadlines that most often concern a small company:

Since when

What applies

2. 2. 2025

Prohibited practices (Art. 5) and the AI literacy obligation (Art. 4)

2. 8. 2026

Transparency in AI (Art. 50), for example for chatbots and AI-generated content

2. 12. 2027

High-risk systems under Annex III, for example AI used in recruitment and employee evaluation (postponed from 2 August 2026)

The deadlines come from the aiacto.eu analysis. The policy will help you prepare for all three.

What the policy must contain

A good policy for a small company has 10 sections. Each is just a few sentences, not a legal treatise.

  1. Purpose and scope. Who it applies to (employees, part-timers, external contractors) and which tools it covers (chatbots, AI in email, translators, image generators).
  2. Approved tools. A list of the tools employees may use and the versions that are permitted. Ban personal free accounts for company data.
  3. Prohibited inputs. What must not be entered into AI: personal data of clients and employees, trade secrets, contracts, passwords, non-public financial data.
  4. Personal data protection (GDPR). When entering personal data is exceptionally permitted, on what legal basis, and that a data processing agreement must be concluded with the tool provider.
  5. Output review. AI can be wrong. Every output that goes outside the company (to a client, onto the website, into a contract) must be checked by a human.
  6. Prohibited uses. AI must not make decisions about people on its own (recruitment, evaluation, dismissal) and must not be used for prohibited practices under Art. 5 of the AI Act.
  7. Transparency. A chatbot on the website must state that it is an AI. AI-generated content that could be misleading must be labelled.
  8. Intellectual property. Rules for AI-generated texts and images: the tool's licences, third-party copyright, citing sources.
  9. Responsibilities and incidents. Who approves new tools, who the contact person is, and whom an employee reports to when they have entered something they should not have.
  10. Training, records and review. Who is trained and when (evidence for Art. 4), how violations are handled, and when the policy is reviewed, at least once a year.

Policy template

Nižšie je skrátený vzor. Hranaté zátvorky doplňte údajmi svojej firmy a zoznam nástrojov upravte podľa toho, čo skutočne používate.

AI Usage Policy of [Company Name]

Effective from: [date] · Approved by: [name, position] · Responsible person: [name]

  1. Purpose and scope

This policy sets out the rules for using artificial intelligence tools when working for [Company Name]. It applies to all employees, part-time workers and external contractors.

  1. Approved tools

You may only use the following tools for work: [e.g. ChatGPT Team, Microsoft Copilot]. A new tool may be used only after approval by [name]. Personal accounts for work tasks are prohibited.

  1. What you must not enter into AI

Do not enter personal data of clients and employees, trade secrets, contracts, passwords or non-public financial data into AI tools. If the data is necessary, anonymise it first.

  1. Personal data and GDPR

Personal data may be entered into an approved tool only exceptionally, with the consent of [name], and only if a legal basis and a data processing agreement with the tool provider are in place.

  1. Output review

AI output is a draft, not a finished result. Before it is sent to a client, published or used in a contract, the responsible employee must check it, especially facts, figures and citations.

  1. Prohibited uses

AI must not be used for making autonomous decisions about people (recruitment, evaluation, remuneration, termination of employment), for manipulating individuals, or for creating misleading content.

  1. Transparency

If a client communicates with our chatbot, we inform them that they are communicating with an AI. We label AI-generated content that could be misleading.

  1. Intellectual property

When using AI outputs, we comply with the tool’s licence terms and third-party copyright. We verify sources.

  1. Incidenty

If an employee enters data into an AI that they should not have, they must report it immediately, at the latest within [24] hours, to [name, email]. Timely reporting is not grounds for a sanction.

  1. Training and review

Every employee completes an introductory training on the use of AI, and a refresher training after each change to this policy. Attendance is recorded. We review the policy at least once a year.

This template is for information purposes only and does not replace legal advice. Adapt it to your company before use.

Want the complete version? The extended 13-page policy contains 11 articles, an employee checklist, an incident record, a training log and an acknowledgement confirmation. Enter your email and we will send it to you in Word so you can edit it right away. You will find the form below.

How to implement the policy in your company: 5 steps
  1. Find out what is already in use. Ask your team which AI tools they use, including through personal accounts. It is often surprisingly many.
  2. Choose approved tools. Prefer company accounts and review the provider's terms: whether inputs are used to train models and where the data is stored.
  3. Adapt the template and appoint a responsible person. Fill in the list of tools, the names and the deadline for reporting incidents.
  4. Train your team and record it. A short training session with an attendance sheet and a date is enough. This record is your evidence for Art. 4 of the AI Act.
  5. Keep records and set up a review. Maintain a list of the tools in use and update your records of processing activities, if you keep them. Review the policy at least once a year.

Most common mistakes

  • General statements with no specifics. The sentence "use AI responsibly" tells nobody what is allowed. You need a list of tools and specific prohibitions.
  • Banning everything. A blanket ban usually ends with employees using AI secretly through personal accounts. It is better to allow approved tools with clear rules.
  • No proof of training. Even a well-trained team is a problem during an inspection if the company cannot prove it.
  • Missing contract with the provider. If you enter personal data into a tool, you also need a data processing agreement.
  • The policy is written once and forgotten. The AI Act keeps changing; in 2026 it was amended by the Digital Omnibus. Set a review date.
  • Copying someone else's template without changes. A policy for a bank does not suit an e-shop with 10 people. Adapt it to the tools and data you actually work with.

Frequently asked questions

Does a small company also need an AI usage policy?

The AI Act does not explicitly require a document called a policy. However, Article 4 requires measures to promote AI literacy, and evidence is expected during an inspection. A policy together with a training record is the simplest way to have it.

How many pages should it have?

2 to 4 pages is enough. What matters is that employees read it and know how to act on it.

Can employees use ChatGPT?

Yes, if the tool is on the approved list and the employee follows the rules on what may be entered into it.

Does the policy also apply to part-time workers and external contractors?

It should. State this in the section on scope.

How often should it be updated?

Aspoň raz ročne a vždy, keď zavediete nový nástroj alebo sa zmení legislatíva.

Does this template replace legal advice?

No. The template is for information purposes only. In specific situations, such as AI in recruitment or the processing of sensitive data, consult a lawyer.

As of 9 November 2026. The legal information is provided for information purposes only and does not replace legal advice.

Download the extended AI Usage Policy for your company

A ready-made 13-page document that you only need to fill in with your company’s details. Enter your email and we will send it to your inbox.

- 11 policy articles ready to edit
- Employee checklist
- Incident record
- Training log and acknowledgement confirmation

Scroll to Top