AI usage policy for your company:
How to build it from within

An AI policy that no one reads is paper. A policy that the team co-constructs becomes a criteria. Ethiceye helps companies build the second.

The difference between a policy that is archived and one that is used

Generic policy (the one that doesn't work)

The legal wording and the download by email

No one knows what he says exactly

Filed when a new tool appears

Covers the file

Generates resistance (They control us)

Politics co-constructed with Ethiceye

It is built by the team in 5 sessions

Each person can explain it with their words

It has an integrated update protocol

Change how decisions are made

Generates property (We decided this)

What does your policy have at the end of the process?

Inventory of approved tools
What does it answer?

What can we use?

concrete example

ChatGPT Enterprise with DPA: Yes.
ChatGPT Plus Personal: Only for internal drafts without customer data.

Data classification
What does it answer?

With what data?

concrete example

Customer data Individuals: Prohibited in external tools without DPA.
Non-personal internal data: Approved.

Use Cases by Role
What does it answer?

For what and who?

concrete example

Marketing can use AI for drafts. HR cannot use IA to filter CVS without human supervision.

Approval Protocol
What does it answer?

How is something new approved?

concrete example

Application to IT → Evaluation in 5 days → Approval or proposed alternative.

human supervision
What does it answer?

Who is reviewing?

concrete example

No AI-generated content for clients is published without human review.

Update
What does it answer?

When is it reviewed?

concrete example

Semi-annual review. Automatic when there is relevant regulatory change (AI Act, GDPR).

Frequently Asked Questions

How to create an AI usage policy for my company?

An effective AI usage policy requires three previous steps: Inventory of which AI tools use the equipment, data classification by sensitivity, and map of current uses. Without that diagnosis, any policy will be generic and not used. The Ethiceye process builds the policy with the team, not for the team.

The initial diagnosis can be made in 1–2 weeks. A complete operational policy, in another 3–4 weeks. The time depends on the size of the equipment and the complexity of the use of AI in the organization.

A Compliance Policy is written by the Legal Department and is published. A criteria policy is co-constructed with the team, part of the existing values and generates criteria that the team uses autonomously. The difference is in whether the team can explain politics in their own words.

Doesn't your company still have AI policy?

Every week that passes, your team makes AI decisions without criteria. Building it takes 5 weeks. Call to start, 30 minutes.

Contact with Raquel López