How to Use AI to Write a Text: A Practical Guide
Using AI to write a text is straightforward: give a precise context to a tool like ChatGPT or Gemini, write a clear instruction (called a prompt), review the output, adjust it to your voice, and publish. Five steps. No magic. But a lot of discipline.
The problem is that most executives and teams stop at step two. They type a vague sentence, get a generic text, and conclude that AI is useless. Or worse: they publish the text as-is, without reviewing it, and it shows.
Here is how to do it differently.
Step 1: Define What You Actually Want
Before opening any tool, answer three questions. Who is this text for? What is the goal (inform, persuade, recruit, sell)? What tone do you want (formal, direct, warm)?
If you cannot answer these three questions, the AI cannot either. It will produce something plausible but hollow.
An HR director writing a job posting gives a different instruction than a marketing director writing a newsletter. Context changes everything.
Step 2: Build a Prompt That Works for You
A good prompt is a work instruction, not a vague question. It contains four elements: the role you assign to the AI, the context of the situation, the specific task, and the expected format.
Concrete example. Instead of typing “Write a job posting for an accountant”, write: “You are an HR director at an industrial SME in Morocco. Write a job posting for a senior accountant, minimum 8 years of experience, bilingual French-Arabic. Direct and professional tone. Maximum 250 words.”
The difference in output is immediate.
For enterprise use cases, I recommend building an internal library of validated prompts. Each team (HR, marketing, legal, sales) has its own templates. This avoids reinventing the wheel each time and ensures tonal consistency.
This is exactly what I cover in my 2-to-3-week AI Governance Sprint, which helps leadership teams structure their AI practices without improvising. Learn more about my services.
Step 3: Choose the Right Tool for Your Use Case
Not all tools are equal for all contexts. Here is what I observe in the projects I work on.
ChatGPT (OpenAI) remains the most versatile. Long-form writing, reformulation, document summarization, complex question answering. It is the reference tool to start with.
Gemini (Google) integrates well into Google Workspace environments. Maroc Cloud announced the launch of Gemini Enterprise in Morocco, opening concrete possibilities for companies that want to manage AI usage within a secure environment.
Jasper and Copy.ai are oriented toward marketing and commercial content. Useful if your team produces many short texts (ads, product descriptions, emails).
Tool selection also depends on your data policy. A recent signal from cio-mag.com indicates that 42% of enterprise AI users in Morocco import complete documents into uncontrolled external tools, which raises a confidentiality risk. Before choosing a tool, verify where your data goes.
I detailed the selection criteria in my comparison of the best AI tools for businesses in 2026.
Step 4: Review, Adjust, Own the Text
AI produces a first draft. Not a final text.
What AI cannot do: know your company culture, your client relationships, the nuances of your sector, your way of expressing yourself. It can imitate a style if you give it examples, but it does not know you.
Review is the moment where you transform a generated text into a signed text. Check the facts, adjust the tone, remove the hollow phrases AI tends to favor.
An unreviewed AI text is detectable. And it damages your credibility.
Step 5: Integrate AI into a Process, Not an Individual Habit
The real gain does not come from one employee using ChatGPT on their own. It comes from a team with clear rules: which tools, for which uses, with which verification steps.
What I observe with my clients: the teams that progress fastest are those that have formalized their practices. They have prompt templates, validation processes, and a policy on what not to share with an external tool.
This is what AI literacy means in practice. Not a one-day seminar. A structured daily practice.
If you want to structure this approach in your organization, request a free diagnostic. We look together at where you stand and what is blocking progress.
Pitfalls to Avoid
Giving vague instructions. AI is not a mind reader. The more precise you are, the better the result.
Using AI for sensitive texts without human validation. Contracts, crisis communications, delicate HR messages: AI can help structure, but the final decision belongs to a human.
Ignoring the data question. Before importing a confidential document into an external tool, read the terms of service. Or choose an enterprise solution with contractual guarantees.
Neglecting team skill development. AI does not replace judgment. It amplifies it. A team member who cannot write well will not produce better texts with AI. They will produce mediocre texts faster. Skill development remains essential, as I explained in my analysis of AI’s impact on human resources.
What You Can Expect
An executive or team that applies these five steps produces texts faster, with fewer back-and-forth cycles, and improved tonal consistency. The gain is not in the volume of text produced. It is in the time freed up for what genuinely requires your judgment.
FAQ
Can AI be used to write legal or contractual texts?
AI can help structure a document or draft a first version. But any text with legal implications must be reviewed and validated by a legal professional. AI makes factual errors and does not know the specifics of your jurisdiction.
What is the difference between ChatGPT and Gemini for writing?
ChatGPT is generally more flexible for long texts and complex instructions. Gemini integrates better into the Google ecosystem (Docs, Gmail, Sheets). The choice depends on your existing tools and data policy.
Can AI write in multiple languages?
Yes. Large models like ChatGPT or Gemini work in French, Arabic, English, Spanish, and other languages. Quality varies by language: French and English generally yield the best results.
How do I prevent AI texts from sounding generic?
Provide examples of your existing texts in the prompt. Ask the AI to imitate your style. Review systematically and remove generic phrases. The more you personalize the instruction, the more the result sounds like your voice.
Is AI writing suitable for SMEs?
Yes, and often more so than for large organizations, because SMEs have fewer writing resources. An SME executive can produce job postings, commercial emails, and communication content without hiring a full-time writer. The condition: take the time to formulate instructions well and review the results.