A policy decision

Ban AI tools at work, or teach people to use them well

Your CISO just found out three teams are already using AI tools nobody approved. Now there is a decision to make: shut the tools down, or bring the use into the open and guide it.

The ban

Fast: a no-AI-tools policy goes out in one email and applies from day one.
Zero procurement: no new register, no rollout, no budget line to defend to the board.
On paper, risk drops to zero the moment the policy lands, because officially nothing is in use.

Guided use

The /start intake asks which AI systems are already in use, by which roles, and builds a role-based route from real usage, not a guess.
Every assignment is recorded with the answers given and the rule version at the time, so it can be replayed later.
The AI register connects declared tools to the roles that use them, so tools stay visible instead of moving to private devices.
Depth follows risk, not job title, across 15 sector-specific scenario sets, so training matches what a role actually does with AI.
The organisation-level evidence file is exportable by org admins, showing measures taken, not just a policy that was signed once.

See the public log of legal changes

A ban does not stop AI use, it stops seeing it

Employees who solved a task with a chatbot last week don't stop because a policy says no. They use their own phone, a personal account, a browser tab nobody audits. The tools disappear from the register, not from the workday. LearnWize's /start intake asks each person which AI systems they actually use, by sector and role, and builds their route from that answer, not from the policy. What is declared becomes visible: which tools, which roles, which risk. What is banned on paper but used in private stays exactly where a ban put it: out of sight.

When a ban is simply the right call

Sometimes the ban is correct, not because it looks safer, but because it is. If a specific tool cannot yet be used safely in a specific context, for example a role handling data or decisions a provider has not yet secured for that use, restricting it while the risk is assessed is a legitimate choice. The difference is that it should be a decision, not a default: name the tool, the reason, and a review date, and revisit it once the risk is actually assessed. A ban without a review date usually outlives the reason for it.

Article 4 asks for measures, not a policy document

Since 27 July 2026, article 4 of the EU AI Act asks providers and deployers to take measures that support AI literacy, taking their own context into account. It does not require guaranteeing a specific level of skill, and a ban is not a substitute for it. If people in an organisation are in fact using AI, an internal ban does not remove that obligation: the article looks at what is actually used, not what is officially allowed. Guided use keeps a record of the measures taken, per role, with the rule version behind each assignment, so there is something to show beyond the policy itself.

See what this looks like for your organisation

The 5-minute scan maps your AI use to the roles around it and shows where your evidence is thin.

AI Act evidence routes

Pick the route that matches your AI Act evidence question.

Focused pages per question: from mandatory Article 4 literacy and training records to sector evidence and an audit-ready dossier.

ChatGPT AI literacy

Help teams use generative AI responsibly with data, source, bias, and output checks.

View route

Copilot AI literacy

A role-based literacy route for teams rolling out Microsoft Copilot or similar workplace AI.

View route

AI literacy for employees

AI literacy for employees: learn safe AI use, recognize risks, test understanding, and build certificates and training records.

View route

Mandatory AI literacy

What Article 4 asks from teams that use, buy, manage, or oversee AI systems.

View route

Guided use

The /start intake asks which AI systems are already in use, by which roles, and builds a role-based route from real usage, not a guess.

Every assignment is recorded with the answers given and the rule version at the time, so it can be replayed later.

The AI register connects declared tools to the roles that use them, so tools stay visible instead of moving to private devices.

Depth follows risk, not job title, across 15 sector-specific scenario sets, so training matches what a role actually does with AI.

The organisation-level evidence file is exportable by org admins, showing measures taken, not just a policy that was signed once.

See the public log of legal changes

A ban does not stop AI use, it stops seeing it

When a ban is simply the right call

Article 4 asks for measures, not a policy document