AI citation study: Websites AI cites most in Human Resources

Two audiences ask HR questions. Candidates and freelancers want to know "how does portage salarial work", "how to find a work-study contract in marketing", "which temp agency hires warehouse staff in Lille" or "is a career coach worth the money". Employers ask "how to recruit a developer quickly", "best executive search firm for a CFO" or "what does a staffing agency cost compared with a direct hire".

Many of these prompts are legal questions in disguise. Behind "can I chain two fixed-term contracts" or "what is the minimum pay in portage salarial" lies a rule from the labour code. The AI must get the rule right before recommending anyone, and that changes the sources it selects.

Students form a third audience: internships, apprenticeships and work-study programmes generate a steady stream of prompts that mix career advice, school choice and employment rules.

In French-language answers, government websites carry much of the factual load: Légifrance for legal texts, the public service portals, the labour ministry, France Travail (the public employment service), URSSAF for social contributions, and the official portals for students and work-study contracts. For English prompts, models turn to the equivalent national authorities of the country concerned.

The second layer is made of job platforms and professional networks. LinkedIn, Indeed, Glassdoor, Welcome to the Jungle and HelloWork are cited for job offers, salary ranges and employer reputation. Freelance marketplaces such as Malt appear when the question is about independent work.

The third layer is the specialists: recruitment firms, umbrella employment companies and their professional federations, schools that run work-study programmes. They are cited when their pages compare offers, detail fees or answer a niche question better than a general portal.

Labour law questions carry risk, so AI models prefer sources that look accountable: a named author, an update date, references to the applicable rules. An HR company whose guides quote the law correctly and show who wrote them has a better chance of being cited in the explanation, not only in a list of providers.

Comparison prompts are the other battleground. Questions like "best portage salarial company" or "top recruitment agencies for engineers" produce lists, and models build them from comparison articles, review platforms and provider pages that state fees, guarantees and specialisations. Vague marketing pages give the model nothing to compare.

GEO (Generative Engine Optimization) for an HR business therefore has two goals: being the source that explains the rule, and being one of the names in the list.