Tell us about your project
Two days of training in August 2025.
Assistants in place by January 2026.
Fondation Cancer Luxembourg entrusted us with two consecutive assignments: training its teams in generative AI, on 25 and 26 August 2025, then gathering the needs of each department and configuring in its ChatGPT subscription the assistants that meet them, between December 2025 and January 2026. This case study describes the method and what was put in place, without figures: no result has been measured at this stage.
Client: Fondation Cancer Luxembourg, cancer information, prevention, support and research, Luxembourg.
Assignment: in-company AI training for the teams, then needs assessment department by department and configuration of assistants in ChatGPT.
Period: training on 25 and 26 August 2025; assistants set up between December 2025 and January 2026.
AIxH services: Custom AI training, ChatGPT and Claude integration and AI audit & integration.
Starting point
Fondation Cancer informs the public, runs prevention campaigns, supports people affected by cancer and funds research. Its teams produce content in several languages, answer many enquiries and organise fundraising events.
The Foundation had a ChatGPT subscription. For it to serve every team in the same way, each first had to know what the tool can do, then have its recurring tasks installed in it as shared assistants.
Method
Training
On 25 and 26 August 2025, the teams were trained in generative AI as part of our custom training courses: how the tools work, usage rules, exercises on their own content.
Needs
After the training, each department described its recurring tasks, its reference documents and the expected result. Each need retained became an assistant to build, with a time estimate for gathering the information, building the document base, writing the instructions, testing and presenting.
Build
Between December 2025 and January 2026, the assistants were created in the Foundation's ChatGPT subscription, tested on real cases and presented to the teams that use them. Each assistant carries its instructions, its reference documents and its limits.
Discover our ChatGPT and Claude integration for your teams →
Assistants
Writing and translation
Article writing by topic, translation into the public's languages, spelling and grammar proofreading, social media copy, accessible text and image descriptions.
Replies
Replies to recurring emails related to a fundraising event and to fundraising drives, prepared from validated answers, then reviewed before sending.
Analysis and monitoring
An annual report made searchable, help with finding ideas and structuring content, preparation of pictograms, search optimisation of content, monitoring of a topic.
We publish neither the number of assistants, nor the time spent, nor the Foundation's data: this case study describes the method. No result has been measured at this stage, and we write none.
Contact
Tell us the tool in place, ChatGPT or Claude, the departments concerned and the tasks that come back every week. A first conversation is enough to scope the needs assessment.
Talk about your ChatGPT or Claude deployment →