AI candidate evaluationOctopyd vs Greenhouse
Greenhouse Talent Matching compares résumé-derived skills, titles, experience, and industry against criteria the hiring team defines and weights. Octopyd’s Ivy evaluates the evidence behind each candidate’s experience, learns from the hiring manager’s reviews, adapts the criteria and weighting, and re-evaluates the candidate pool.
| AI evaluation capability | Octopyd Ivy | Greenhouse | What Greenhouse documents |
|---|---|---|---|
| AI-assisted résumé-to-job matching | Talent Matching compares candidate résumés with user-defined calibration criteria. Talent Matching ↗ | ||
| Recruiter-defined evaluation criteria | Recruiters define the criteria used to calculate candidate match strength. Talent Matching ↗ | ||
| Weighted evaluation criteria | Users can assign importance to selected skills and other calibration fields. Talent Matching ↗ | ||
| Semantic matching beyond exact keywords | Greenhouse uses embedding representations to match semantically related skills and job titles. Data Processing FAQ ↗ | ||
| Human-readable explanation of the match | Greenhouse shows matched experience, industry, skills, and exact or similar résumé terms. Talent Matching ↗ | ||
| Manual override of the AI result | A recruiter can manually override a match category after reviewing the résumé. Talent Matching ↗ | ||
| Calibration version history | Greenhouse can record who changed the calibration, when it changed, and which version was active. Talent Matching ↗ | ||
| Re-evaluation after a user manually changes the calibration | Greenhouse recalculates current candidates after a user edits or reapplies a calibration. Talent Matching ↗ | ||
| Automatically learns from hiring-manager reviews | Greenhouse states its matching algorithm currently uses résumé data and does not access internal Greenhouse data sources. Data Processing FAQ ↗ | ||
| Learns from advance, hold, and reject decisions | Greenhouse requires a human to advance or reject each candidate and does not document using those decisions to update the calibration automatically. Data Processing FAQ ↗ | ||
| Adaptively changes evaluation criteria | Criteria can be manually edited, but automatic criterion changes based on hiring-manager review patterns are not documented. Talent Matching ↗ | ||
| Adaptively changes criterion weights | Users manually assign and change criterion importance; automatic weight learning from candidate reviews is not documented. Talent Matching ↗ | ||
| Re-evaluates the pool after each adaptive learning cycle | Greenhouse re-evaluates after a user manually updates the calibration, not after an automatically learned feedback cycle. Talent Matching ↗ | ||
| Identifies the hiring manager's implicit priorities within about three review cycles | Greenhouse does not document an adaptive convergence process that infers unstated preferences from repeated hiring-manager decisions. Data Processing FAQ ↗ | ||
| Converts the full job context into detailed evaluation criteria | Greenhouse can generate a calibration from the job description and scorecard attributes, but recommends focusing on four to six key skills, with optional experience, industry, and title criteria. Talent Matching ↗ | ||
| Evaluates every candidate against every detailed criterion | Greenhouse assigns an overall match category based on selected calibration attributes rather than a full criterion-by-criterion contextual evaluation. Talent Matching ↗ | ||
| Criterion-level score, evidence, reasoning, strengths, and concerns | Greenhouse shows matched skills and supporting résumé terms, but does not document a separate contextual score and analysis for every criterion. April 2026 release notes ↗ | ||
| Evaluates demonstrated capability when the exact skill term is absent | Greenhouse supports semantic relationships between terms, but its documented processing is restricted to extracted skills, titles, dates, employers, experience, and industry. Data Processing FAQ ↗ | ||
| Distinguishes demonstrated experience from keyword stuffing | Greenhouse highlights exact and semantically similar résumé terms; it does not document checking whether each listed skill is supported by project-level evidence. Data Processing FAQ ↗ | ||
| Evaluates skill depth: exposure, execution, design, ownership, or expertise | Greenhouse's documented matching attributes do not include depth of application or level of ownership. Data Processing FAQ ↗ | ||
| Evaluates personal ownership and decision authority | Greenhouse does not document extracting or scoring personal ownership, accountability, or decision authority. Data Processing FAQ ↗ | ||
| Evaluates seniority beyond job title and years of experience | Greenhouse processes job titles and years of experience, but does not document evaluating seniority through autonomy, complexity, scope, or organizational impact. Data Processing FAQ ↗ | ||
| Evaluates career progression across roles | Greenhouse extracts employment dates and titles, but does not document assessing progression, stagnation, or declining responsibility. Data Processing FAQ ↗ | ||
| Evaluates scale, complexity, and relevance to the actual business problem | Team size, system scale, decision complexity, company stage, and business context are not among Greenhouse's documented Talent Matching attributes. Data Processing FAQ ↗ | ||
| Separates direct evidence, inference, and interview-validation needs | Greenhouse provides a short match explanation, but does not document separating supported facts from inference or generating candidate-specific validation questions. Talent Matching ↗ | ||
| Independent multi-model evaluation and cross-verification | Greenhouse uses multiple specialized models for extraction tasks, but does not document running the complete evaluation independently across several models and comparing the conclusions. Data Processing FAQ ↗ | ||
| Evaluates internal consistency of résumé claims | Talent Matching is documented as extracting and matching selected résumé attributes, not analyzing contradictions among titles, dates, responsibilities, skills, and accomplishments. Data Processing FAQ ↗ | ||
| Checks relevant claims against available external professional evidence | Greenhouse states Talent Matching currently uses candidate-submitted résumé data and does not access internal sources; external professional corroboration is not documented. Data Processing FAQ ↗ | ||
| Candidate Consistency Score for résumé-claim credibility | Greenhouse does not document a score measuring whether professional claims are internally coherent and supported by available evidence. Data Processing FAQ ↗ | ||
| Application metadata fraud signals | Coming soon | Greenhouse Fraud Detection separately analyzes phone, email, IP address, and location signals. Greenhouse states this fraud feature does not use AI. Real Talent ↗ | |
| Integrated identity verification | Coming soon | Greenhouse offers CLEAR identity verification to confirm that an applicant is who they claim to be; it is not employment-history or résumé-claim verification. Real Talent ↗ |
Where Greenhouse supports a capability, the linked page directly confirms it. Where it does not, the link establishes the narrower documented workflow or data scope. Greenhouse does not publish a page explicitly listing every capability it does not provide, so absence claims are limited to the scope of its documented native product. Greenhouse capabilities and documentation verified July 2026; features vary by plan and change over time.
A note on adaptive evaluation
Greenhouse supports manual recalibration: a user can change the criteria or weights, and Greenhouse re-scores the current review pool. Ivy supports adaptive re-evaluation: it observes the hiring manager’s reviews, identifies repeatable differences between the initial rubric and their actual preferences, updates the criteria and weighting function, and re-evaluates the pool. Within about three review cycles, Ivy is designed to align its recommendations with what the hiring manager truly values.
A note on résumé integrity
Greenhouse provides separate fraud and identity controls that can flag suspicious application metadata and verify that an applicant is a real person. Ivy’s Consistency Score addresses a different question: whether a candidate’s professional claims are internally coherent and supported by the available evidence, across titles, dates, responsibilities, claimed skills, ownership, seniority, and accomplishments, flagging areas to validate during interviews. A Consistency Score is a validation and risk signal, not proof that a candidate is dishonest; the absence of external corroboration should not by itself be treated as evidence that a private or confidential accomplishment is false.
The bottom line: Greenhouse helps recruiters prioritize résumés against a manually defined calibration. Octopyd’s Ivy develops and continuously improves the evaluation framework, learns from hiring-manager decisions, evaluates what each candidate actually demonstrated, and flags claims that require deeper validation.