The result to aim for
Air Liquide has put online, as the first product of its Hub LLM, a multilingual chatbot that answers questions from employees of its shared financial services about their processes and procedures.
What was done
After LeMagIT (30 April 2026), the group first deployed Gemini for Workspace to its 65,000 employees and then relied on the French publisher Prism.ai's platform for more specific assistants. The first is addressed to the Global Business Services, a thousand people in several countries, who must follow the same financial processes described in multiple documents. the chatbot collects these sources in one channel and has been deployed in several languages. According to the Chief Data & AI officer, Fabien Mangeant, the training of newcomers took six months and required more than three. The next step planned is automation by treatment agents related to ERP; he estimates that an agent represents two to three months of work, from design to implementation.
Before you start
Tool cited by the source: Prisme.ai; other tools cited: Gemini. Gather the information the prompt asks for. Strip confidential data that is not needed and check your tool’s access terms. Features and plans can change.
The method, in three steps.
Gather the context
Gather the information described in the prompt.
Adapt the prompt
Copy the prompt below, replace the bracketed items, then paste it into Gemini.
Reread before using
Reread the result, correct it if needed and keep the final decision. Caveats are detailed below.
Your starting point.
Replace the bracketed items, then adapt the proposal to your situation. This prompt is written by the iadujour desk It is not a quotation from the source.
You're the integration assistant of our service. You answer only from the procedures provided: Newcomer question: 1. Answer in numbered steps. 2. List the document and the reference section. 3. If the procedure differs according to the country or entity, ask which one before responding. If nothing covers the issue, refer to the process's reference.
What deserves your attention
Proof level: Testimony published by an identified source. Source checked on 10 November 2026. Use described by the source, not tested by the desk. Original summary written by iadujour.fr: figures and claims are the source’s; the prompt is a desk proposal, not a quotation.
Useful skills
Official skills on the same task, paired by the desk: they are not part of the described case.
- Deploy several services in the same Vercel projectOpenAI · Vercel Services allows to deploy several services within the same Vercel project.
Comments.
A field note, a question, a precision on this use case? Comments are reviewed before publication.
Published comments
Loading comments…
Leave a comment
Moderation charter. Every comment is reviewed before publication. We refuse hate speech, insulting or discriminatory language, advertising, spam and promotional links, off-topic posts and other people’s personal data.