Survey conducted in early 2025 among 5,760 households. Source: ANRT, enquête TIC 2024-2025.
Start with the decision
A useful approach defines the work, the people involved, the current evidence and the result to improve. Technology is assessed after the process, information and human responsibilities are understood.
Diagnose
Observe work, handovers, information needs and sources of delay.
Write the assumption, evidence required and person responsible for the next decision. This makes the work reviewable and prevents enthusiasm for a tool from replacing analysis.
Sequence
Balance quick improvements, structural projects and dependencies.
Use a short scale and a written justification. A score without context creates false precision; the discussion and the evidence are more valuable than the total.
Adopt
Give teams roles, training, support and measurable outcomes.
Define a review date and the conditions for continuing, changing or stopping. A disciplined stop is a valid result when the expected value, quality or control cannot be demonstrated.
Decision table
| Dimension | Practical question |
|---|---|
| Diagnose | Observe work, handovers, information needs and sources of delay. |
| Sequence | Balance quick improvements, structural projects and dependencies. |
| Adopt | Give teams roles, training, support and measurable outcomes. |
Moroccan context and sources
The Moroccan CNDP states that AI processing involving personal data is governed by Law 09-08. Morocco AI 2030 also places AI within a national agenda for sovereignty, trust, innovation and adoption.
Work with Consultor
Consultor connects the method to an assignment. Monitor contributes when monitoring, information or reputation is involved.
How to apply this guide in a real organisation
Start by naming the decision, the people affected, the current process, and the evidence already available. Separate confirmed facts from assumptions. Then identify the business owner, the people who can validate data and risk, and the users who understand day-to-day exceptions. This prevents a technology discussion from replacing the operating question.
Use a small number of explicit criteria: expected value, data readiness, feasibility, consequence of error, human review, cost, and adoption effort. Record the justification behind every assessment. A high score must never hide a blocking legal, security, or operational issue. The result should support a choice to proceed, revise, postpone, or stop.
Evidence and review checklist
- Document the baseline before changing the process.
- Keep the source, owner, and limits of every important claim.
- Define who approves outputs and handles exceptions.
- Test with representative examples rather than a polished demonstration.
- Set a review date and measurable continue-or-stop conditions.
Consultor connects this method with AI strategy, digital transformation, governance, and adoption in Morocco. Monitor contributes when the decision depends on media monitoring, strategic intelligence, reputation, or information signals. The final deliverable should make responsibilities and unresolved questions visible to leadership.
Questions fréquentes
What is the first step for this work?
Start with the business decision, current process, people involved and evidence available before selecting a tool or solution.
Who should own the decision?
A named business owner should be accountable, with input from users and the legal, data, security or technology functions required by the context.
When should the approach be reviewed?
Review it after every material experiment and whenever the process, data, tool, risk or organisational priority changes.