The revolution is underway! The number of artificial intelligence users has increased by 60 % in one year. In companies, where employees used to use AI without talking to their bosses, they are now being encouraged to do so by their superiors. This change in mentality has taken place in record time, given that ChatGPT was launched just three years ago. So now is the time to get started. To be an assistant or secretary today, you need to know how to use AI and be aware of all the pitfalls... Which tools should you use? For what purposes? Élisabeth Duverney-Prêt, a specialist in the assistant and secretarial professions, takes a closer look.

All the IT giants have got involved. Following the success of ChatGPT, new names have appeared on the market: Gemini, Claude, Copilot, Mistral, DALL-E, Midjourney... Each of these AIs has its own specific features.
Training
Generative artificial intelligence, from ChatGPT to MultiGPT
Key points:
- Practical uses ofGenerative AI
- Company professions affected
- Different solutions on the market
Example of practical work: study of a use case "how can generative AI be used in practice within your company"?
Participants speak out:
"Fun training easily accessible to all levels. Practical exercises with real impact. I was very sceptical about AI when I came and I came away from the course with the knowledge to better understand, use and criticise the tool."
"A very good approach to discovering this vast field of generative AI. Excellent exchanges with the participants, who were naturally motivated by the quality of the trainer's performance. Good balance between practical cases, theory and discussions. Especially in 2 days."
AI for greater efficiency
For assistants and secretaries, AI has rapidly become an indispensable partner for a more efficient and effective better time management and increased productivity. Text, image and video creation, table generation, meeting reporting... Everything is now possible.
"There was nothing more tedious than looking for precise data in a file. I was wasting an incredible amount of time. With generative AI, document analysis has become very fast and I can ask it all the questions I need for the tasks I'm working on, so I can move forward more quickly", explains Aurélie, an assistant in a services company in the Paris region.
Those who took the plunge from the outset have become true experts in the field, freeing up time to more creativity. It's a leap forward that's much appreciated by their employers! It's enough to motivate all employees to take the plunge.
On a day-to-day basis, generative AI has thus become a real asset.
AI has become a particularly useful medium for creating content: writing emails, letters, reports or presentations.
It can also be used to generate text for slides, create summary sheets or even write LinkedIn posts based on a given piece of content.
You can also ask it to generate an activity report based on raw data. All spelled correctly.
But then, should we fear that in a few years' time AI will simply take the place of assistants? Certainly not!
On the other hand, when it comes to hiring, employers will prefer someone capable of handling AI agilely. So it's high time to get started.
Overview of AI tools
At first sight, it may seem difficult to make sense of the variety of AIs on the market. To help you, here are a few key points to remember.
ChatGPT, a versatile tool
ChatGPT has the advantage of being versatile, thanks to its multitude of functions. You can ask it to generate statistics, forecasts and even recommendations based on different data. It can synthesise notes from a meeting, bring together information scattered across several documents and summarise texts to extract relevant summaries.
From this document of over 130 pages, it is able to extract a two-sentence summary.
But that's not all: ChatGPT can also be used to write content for communication media, generate ideas for topics or improve texts. For meetings, it can help draft detailed agendas and plan the order of topics to be discussed.
For example, for the next trade fair you're organising, let ChatGPT help you come up with ideas for invitations.
Finally, on a day-to-day basis, it allows you to prioritise tasks according to their urgency and importance.
Would you like to learn about best practice in designing a prompt and find out which generative artificial intelligence tool is best suited to your needs? Discover the training programme The 20 best generative artificial intelligence (GAI) solutions for your business.
Gemini, creativity and synthesis
Google Gemini is also a comprehensive tool. It acts as an analyst, helping to spot trends, synthesise information and identify business opportunities, for example. As a creative tool, it can be used to help create images and illustrations for presentations. It can also be used to take notes and write e-mails. The same is true of Copilot, the AI developed by Microsoft.
If you're short of ideas for developing your brand image, these tools can give you valuable advice.
Mistral and Claude, complementary tools
For its part, Mistral is the rising star of the French version of generative AI. Diary management, copywriting and proofreading, information retrieval, communication, event organisation, translation... An AI with all the bells and whistles and a commitment to data confidentiality. Need to free up an hour in your diary? Don't hesitate to ask Mistral to help you.
Claude has interesting analytical and writing capabilities. This AI simplifies and speeds up complex tasks, such as document analysis, by extracting key information from PDF files or large reports in a matter of seconds. Claude can also generate blog content, product descriptions and detailed reports. For project management, it automates the planning and monitoring of tasks.
These are undeniable advantages for assistants, and are also available from Perplexity AI and Llama3.
All of these AIs are particularly effective for text generationBut that doesn't mean we should forget the images. Many specialist AIs have entered the market, but DALL-E and Midjourney remain among the most popular. As with text generation, prompting is of paramount importance here to ensure that the result is exactly what you want.
Training
Midjourney, learn how to generate professional visuals with AI
Key points:
- Master the Midjourney interface and basic settings to generate images efficiently
- Develop techniques for writing prompts to optimise the quality and creativity of the images produced
- Discover and use additional tools to improve, style and optimise Midjourney results
So there you have it, a clearer picture of the benefits of AI for your business.
AI, where to start?
Where can you find the time to get started? Here's a plan to help you get organised:
1er month: discovery
Week 1: Getting to grips with AI tools
- Mistral for preparing meeting summaries or replying to e-mails.
Week 2: testing the automation of repetitive tasks
- ChatGPT for automatic replies to standard emails or Mistral for summarising Codir meetings in a few lines.
Week 3: file organisation and management
- Copilot to organise an event file (guests, diary, etc.) or Claude to generate follow-up emails.
Week 4: creating visual content
- Create professional visuals for presentations or internal media with Midjourney or DALL-E.
2e month: more in-depth
Week 5: using AI to organise an event
- Mistral to write invitations and programmes for an internal event, and DALL-E to create visual invitations.
Week 6: Improving your writing
- ChatGPT for drafting minutes of Codir meetings or emails to teams.
Week 7: data management and reporting
- Copilot to create dashboards in Excel and generate automatic reports from data.
Week 8: Creation of visual aids
- Create visuals for presentations with Midjourney. Create infographics from data you want to communicate at meetings or events.
3e month: master's degree
Week 9: customising the use of tools
- Use Claude to create recurring e-mail templates for the Executive Board or standard documents for events.
Week 10: AI safety
- Ensure that documents containing sensitive data are protected (use passwords on files or check the confidentiality of exchanges via Claude).
Week 11: Practical exercises
- Mistral for analysing a performance report and DALL-E for creating a results infographic to present at a board meeting.
Week 12: learning review
- Evaluate the effectiveness of AI in organising events, managing Codir meetings and writing reports.
The importance of promptness
To use AI properly, you need to be particularly precise in your request. This is called the prompt and consists of detailing the instructions that the AI should follow. The prompt must be clear and unambiguous. It is therefore necessary to use vocabulary adapted to the context and to give as many details as possible using short sentences.
Would you like to learn how to create effective prompts to generate specific responses or complete precise tasks? Discover the training programme Prompt engineering, communicate effectively with artificial intelligence.
Writing a reply to an e-mail: example of a prompt
"I'd like to reply to an e-mail from one of my colleagues. He's asking me for information about the seminar project. I'd like to include a timetable and a reminder of the date of the next team meeting."
The AI makes an initial draft and then invites you to add the various dates. If the tone doesn't suit you, you can adapt it by asking for something more formal, friendly, direct, etc.
Taking minutes at a Teams meeting: example of a prompt
There are three steps involved: first, you need to activate transcription in Teams in order to generate the text automatically. Then you need to download this transcription and submit it to the AI with the following prompt:
"Can you summarise this meeting transcript in 1,500 words? This summary will be sent to all our export customers. I'd like to include the sales trend chart.
But beware: your transcription may be too long. You then have two options: submit it in several parts to the AI or upgrade to the paid versions, which offer more extensive features.
Points to bear in mind
Whatever the result of the query, it's essential to remember that you can't rely on AI-generated text with your eyes closed. Many errors can creep in.
It is essential to question the veracity of the information produced.
So don't hesitate to reread the texts and check the data by cross-referencing other sources. Your reputation is at stake.
Beware also of information overload. With its international network of websites, AI often tends to transcribe all the information it has on a given subject, but some of it will probably not be useful to you. So don't hesitate to reframe your request.
Finally, as generative AI models have been created by human beings, certain bias are sometimes present in the texts produced: discrimination, bias, deliberate omission of information, etc. It's up to you to use your critical faculties to give your texts the value they deserve.
Once you've taken these few points into account, you'll know how to make AI your productivity asset.
Generative AI enables a quantum leap in the management, organisation and creativity of assistants.
Knowing how to use AI, in a measured way, has become an asset that is highly valued by employers, along with the hindsight and critical thinking needed to use it properly. And what about tomorrow? AI is making constant progress. Conversational assistants seem to be gradually gaining the upper hand, but future progress will undoubtedly focus on more precise and collaborative models. It's up to you to jump on the bandwagon now, so you don't get left behind in the months and years to come!