Clara and the shortlist problem
Why ranked shortlists with plain-English fit explanations beat keyword search at the mid-market level.
Shortlists are the choke point of every recruitment process. The brief is rich, the candidates are abundant, the time is short, and the explanation is missing. Keyword search returns a list. A list is not a shortlist.
Clara, the co-pilot inside Astrala Nexus, treats the brief as the first-class artefact. GPT-4o reasoning extracts the operational signal. The Pi emotional intelligence layer scores cultural and relational fit. The LIT framework provides the rubric. The output is ranked, with each rank carrying a plain-English explanation of why the candidate sits where they sit.
The match threshold is published. The science is documented. The operator sees the system, not a black box.
This is what science-grounded matching looks like at the mid-market level. Not a magic trick. A measured rubric, applied consistently, with the explanation visible.