AI match score explainability provides natural-language insights into why a candidate is a strong or weak fit for a role. It breaks down the match score into two core dimensions—Skill Relevance and Experience Relevance—and presents evidence-backed explanations for each, so recruiters can quickly understand the reasoning behind a candidate’s score.

No. The underlying match score calculation remains unchanged. Explainability is a new presentation layer that provides qualitative, human-readable reasoning on top of the existing score. It does not alter, recompute, or replace the match score itself.

No. Explainability uses a separate AI model specifically designed to generate natural-language insights. This model takes into account the existing match score, calibration data, and candidate profile information to produce its explanations, but it operates independently from the scoring algorithm.

Explanations are available in two locations. On the Pipeline page, hovering over the match score displays a compact summary with one-line insights under Skill Relevance and Experience Relevance. On the Candidate Profile page, a dedicated panel shows the full detailed breakdown, including individual hiring criteria, evidence tags, and sentiment indicators.

The feature currently draws from four types of evidence: the candidate’s resume (shown with a document icon), profile data, AI interview feedback (shown with an AI interview icon), and human interview feedback (shown with a calendar-person icon). Each explanation line includes evidence tags so you can trace exactly where the insight comes from.

No. Interview feedback doesn't change the underlying match score. However, interview feedback is used by the explainability feature to generate richer, more complete explanations about a candidate’s fit. The match score and the explanation layer are separate.

Each explanation line includes visual indicators. A positive (upward) icon with a positive color means there is strong supporting evidence for that criterion. A negative (downward) icon with a negative color means evidence is missing or conflicting. A neutral or grey icon indicates that evidence is incomplete or indirect. Evidence source tags (such as resume, interview, or feedback icons) are also displayed and are clickable, linking directly to the underlying source.

These are the two core evaluation dimensions. Skill Relevance assesses how well the candidate’s skills, competencies, and knowledge align with the role requirements, including functional readiness and transferability. Experience Relevance evaluates how closely the candidate’s career trajectory, prior roles, industry background, and seniority level align with what the role expects.

The system identifies up to the top ten most important hiring criteria for each role, drawn from the job description and calibration parameters. This ensures the explanation focuses on what truly matters for the position rather than covering every minor detail.

Currently, evidence tags are not clickable. They serve as visual indicators showing which data source—such as resume, profile, AI interview, or human interview feedback—contributed to each explanation line. This helps you understand the basis of each insight at a glance. Clickable evidence linking to the underlying source material is planned for a future release.

Ideal Candidate data isn't directly used in the explainability output itself. Instead, it plays an indirect but important role: the system analyzes the Ideal Candidate profiles to identify which features (such as skills, titles, and experience attributes) are most relevant for the position. The features where the candidate matches the Ideal Candidate are considered the most important ones, and these same features become the hiring criteria used to generate the explainability insights. Ideal Candidates help determine what matters most for the role, and the explanations are then built around those key criteria.

The explainability feature uses an OpenAI model by default. For customers who have explicitly opted out of OpenAI, an Eightfold self-hosted model is used instead. This means that regardless of your organization’s OpenAI preference, the explainability feature is available to you—either through OpenAI or through Eightfold’s own hosted model.

The explainability feature is available to organizations that are on the refreshed candidate profile experience and whose candidates have match scores for the relevant positions. Please contact your account representative to confirm your organization’s configuration and enablement status.

No. AI match score explainability is currently available only on the recruiter-facing Pipeline and Candidate Profile pages. It is not displayed on the candidate-facing career site.

Yes. AI match score explainability is available for all languages supported by the Eightfold platform. Explanations will be generated based on the available candidate and position data regardless of the language used.

The explainability feature is designed to align its tone and balance with the underlying match score. For a high overall score, you will generally see more positive signals, and for a lower score, more gaps will be highlighted. The explanation is meant to complement—not replace—recruiter judgment. If something seems unexpected, review the linked evidence sources for additional context.

Explanations are cached for performance. When key inputs change—such as the job description, calibration parameters, or the candidate’s resume or profile—the cached explanation is refreshed so that subsequent views reflect the updated information.

Yes. The explanation panel includes a note that insights are evaluated from available evidences and that recruiters should use discretion when reviewing them. The feature is intended as a decision-support tool and should always be used alongside professional judgment.