One analysis at a time
Retrieval is bound to the analysis being viewed.
Forma. separates measurement from generation. Candidate order comes from fixed, inspectable signals; AI helps explain the evidence around that order.
Explore the productThe current default weights are read from Forma.'s scoring configuration.
How closely the candidate's experience aligns with the role's meaning.
Direct evidence of named technologies and requirements.
How much of the role's required skill set is evidenced.
Each stage has a distinct job. The source document and measured components remain available after ranking.
A job description and one or more candidate documents enter the analysis.
PDF, DOCX, XML, and TXT take format-aware extraction paths. OCR can recover scanned PDF text.
Skill aliases and role requirements are structured for a consistent comparison.
Semantic alignment, explicit keyword matches, and required-skill coverage are calculated separately.
Fixed weights combine the three signal scores into a final value.
Candidates are ordered deterministically. Explanations describe the resulting position.
Ask Forma. retrieves indexed evidence from the current analysis for recruiter questions.
AI explains the result.
It does not decide the rank.
Ask Forma. retrieves document chunks from the active analysis and returns source references with its response.
Retrieval is bound to the analysis being viewed.
Indexed chunks supply the answer's source material.
Answers can refer to the deterministic results already calculated.
Source references and excerpts accompany the answer in the product.
Run an analysis and read each candidate through the same measured signals and evidence path.
Start analysis