Case Study: nine AI assistants for a communications team
By the end of the training, all five members of the communications team said they would increase their use of AI
Situation
The communications team of this Geneva public institution in the social sector, anonymised at its own request for confidentiality, produces everything the organisation shows of itself: intranet, website, social media, newsletters, annual report, event materials. Five people, with very different profiles and levels of comfort with AI, and a stake that goes beyond productivity: making an institution known that is still too discreet relative to its impact, and strengthening its employer brand. At the scoping stage, the institution had no policy at all on the use of AI: everyone was experimenting on their own, without a framework and without a method.
Objective
The team had very precise needs, which it listed for us in writing before the training: writing and adapting copy for a given audience and channel, turning texts into easy-to-read language, creating visuals that respect the graphic charter, subtitling videos, and automating the editorial calendar, event checklists and standard replies. The classic risk in this situation: an off-the-shelf course, disconnected from those needs, that impresses on the day and changes nothing the following Monday.
Will increase their use of AI
5 out of 5
AI assistants delivered
9
Format
1.5 days
Documentation handed over
4/4 unanimously
Process
Day 1: laying the foundations
In July 2025, a full day on the fundamentals: an overview of generative AI tools (text, image, audio, video), how a chatbot works and where it stops, the classic mistakes made in a professional setting, and the six-part structure of a prompt. Every example was built on their own reality: the thread running through the day was building, block by block, a complete LinkedIn strategy for their institution. The day alternated theory and practice: a timed 15-minute challenge, a quiz to check what had been learned, and exercises in pairs.
Day 2: turning every need into a tool
A half-day built entirely on the list of needs sent by the head of communications. Every point on her list became a use case worked through in the session: an image generation assistant applying the institution's graphic charter, an easy-to-read assistant, an assistant that turns a job advert into social media posts, a rewriting and simplification assistant, an editorial calendar assistant, an event checklist assistant, a newsletter assistant, the creation of quizzes and animations in the institution's colours, and a bespoke web application to adapt a piece of content to the target audience and the platform.
What the team takes away
The team left with a library of nine documented, ready-to-use assistants, an online resource space gathering every prompt used, and a guide to image generation.
Results
All five participants answered the end-of-training questionnaire. To the question "Will you increase how often you use AI in your work?": five yes out of five.
Likelihood of recommending the training (scale of 0 to 10): four scores of 10 and one of 9.
Overall satisfaction: 4/4 for four participants out of five, 3/4 for the fifth.
Quality and usability of the documentation handed over: 4/4 unanimously.
The only area for improvement raised, both in the questionnaire and by email a few days later, was the length. "I would have loved the training to go on longer." When a team's complaint is that it wants more training, that is the sign of an appetite that has taken hold, not of a need left unmet.
AI adoption is not decreed in a charter, it is built on real tasks. What tipped this team was not the overview of tools: it was seeing its own list of needs, sent a week earlier, turned point by point into assistants they could use that same day. Generic training informs. Training built on your tasks equips. Does your team have its own list of needs? That is exactly where we start.
Testimonials
- Very interesting training that more colleagues could benefit from (a participant)
- Dynamic, competent, pleasant, I recommend him (a participant)
- In the time available, we managed to cover a very wide spectrum and everything I had asked for, and then some! (the head of communications)
Contact us
- Want to upskill your teams?
Let's schedule a meeting to assess your training needs
- Want to identify high-ROI use cases for generative AI?
Let's schedule a strategic call