Where should an administration or company in Morocco start with AI?

In short

Start with governance before the tool: list the AI uses already in place, appoint an owner, set rules on personal data in line with Law 09-08, then choose one or two high-volume, low-risk use cases, with an outcome indicator measured before and after.

1. List existing uses

Staff often already use AI assistants, with no framework. A simple inventory (who, which tool, which data) reveals the immediate risks and the real needs.

2. Set the rules

An appointed owner, data banned from public tools, human validation of outputs, traceability. For personal data, apply Law 09-08: purpose, minimization, informing the people concerned.

3. Choose a first use case

Aim for a repetitive, high-volume task: sorting mail, standard replies, summarizing documents, searching internal texts. Avoid individual decisions about people at first.

4. Measure and extend

Measure lead time or quality before the project, then after three months. Extend only what has proven itself, while training the staff concerned.

Take the AI governance diagnostic →AI governance

Frequently asked questions

Do you need a full AI strategy before starting?

No. Basic rules and a first measured use case are worth more than a strategy that stays on paper; the strategy is built from the results.

Can public AI tools be used with internal data?

Not without rules: personal and confidential data must not be entered, unless a tool is approved and contracted for that use.

How do you measure an organization’s AI maturity?

Across several pillars: vision, governance, data, skills, technology and uses, each scored on a scale of levels.