CVs, cover letters, and generic skills tests are increasingly easy to game. With AI, candidates can optimise language, rehearse answers, and pass abstract screening tools without ever demonstrating real job readiness.
Realistic Job Assessments and (well-designed) work sample tests are different because they are hard to fake with AI.
Both focus on what candidates do, not what they claim:
-
work sample tests require candidates to complete real tasks
-
Realistic Job Assessments place those tasks inside real job scenarios, with trade offs, pressure, and consequences
AI can help someone write about a job. It cannot easily replicate how someone prioritises work, responds to realistic situations, or reacts to the actual conditions of a role.
RJAs go one step further by testing fit and commitment, not just task output. Candidates must engage with the realities of the job including shift patterns, locations, pace, and expectations. This makes RJAs resistant to automation gaming and far more commercially useful than black box screening tools.



