The Biggest Change in Candidate Behaviour Since the Internet

The candidate research journey has quietly shifted upstream. Where employer discovery once began on job boards, search engines and recruiter outreach, it is now increasingly mediated by AI assistants — ChatGPT, Gemini, Claude and Perplexity — that synthesise public signals and return a small set of recommended employers.

This is the most consequential change in candidate behaviour since the consumer internet. For the first time, a candidate can ask one question and receive a curated employer shortlist before ever touching a careers site.

When discovery is mediated, presence is no longer enough. An employer can be highly visible in traditional channels and still be absent from the recommendation set that candidates increasingly encounter first.

This reshapes the economics of employer brand. Reach metrics describe activity. Recommendation environments describe outcome — whether you are surfaced when a candidate is actually deciding where to apply.

Across analysed prompts, employers with similar awareness scores often show very different recommendation patterns. Recognition does not translate cleanly into recommendation. Some well-known organisations are not surfaced at all for category-defining questions; lesser-known competitors with stronger structured signals appear consistently.

• Recommendation visibility varies sharply by engine, prompt category and market context.
• Industry assumptions sometimes substitute for organisation-specific evidence.
• Employers with consistent structured signals are referenced more often than those relying on traditional brand presence alone.

Employer brand teams need a second measurement layer alongside traditional reach and engagement: how the organisation is described and recommended when a candidate asks an AI assistant a real question.

The relevant question is no longer only 'how visible are we?' but 'when a candidate asks an AI assistant about employers in our category, what is observed?'