Purpose

This article helps you understand the AI match score explainability displayed for a resource against a demand, what you see in the UI, and how to interpret it.

Overview

AI match score explainability expands on what makes a resource a strong or weak fit for a demand based on their skills and experience, supported by evidence drawn from their resume, profile, and interview feedback.

The explanation doesn't change the match score. It's a reasoning layer on top of a score that's already been calculated.

Where you see AI match score explainability

Explanations appear wherever a match score already appears in Resource Management:

• Demand pipeline—point at a resource's match score for a compact summary
• Resource profile—the Highlights panel shows the full breakdown for the current demand
• Matched Demands widget on the resource profile—the breakdown for each demand that resource matches
• Sample resources on the demand calibration page

An explanation is available when the resource has a match score for the demand, and your organization has the Resource Management explainability feature enabled.

Open match score explainability in the resource profile

1. Open a demand and select a resource to go to the resource profile.
2. Go to Highlights and find the match score for that demand.
3. Read the panel. It opens with separate Skill relevance and Experience relevance sections.

To see a compact view, point at a resource's match score on the demand pipeline. The card opens on hover and shows a shorter summary for each of the two sections.

What you see in the panel

The panel is organized into two sections.

Skill relevance

Definition: How well the resource's skills, competencies, and knowledge align with what the demand requires.

Includes:

• Functional readiness and what the resource can do
• Breadth and depth across the skills the demand requires
• Transferability of knowledge to the demand

Experience relevance

Definition: How closely the resource's delivery history and working context align with the demand's expectations.

Includes:

• Depth and relevance of prior engagements, roles, and overall background
• Fit with the seniority, industry, and delivery context the demand calls for
• Progression and overall readiness signals

Section details

Each section contains:

• A short summary giving an at-a-glance view of that dimension, which is what the pipeline hover view shows
• A list of criteria lines, each with a natural-language explanation, an icon and color indicating whether the signal is positive, negative, or neutral, and one or more evidence tags showing where the signal came from

Eightfold consolidates explainability into these two categories because they reflect how staffing decisions are actually made: what the resource can do, and where they've done it. The explanation is based on the evidence the system can access, and it should complement your judgment rather than replace it.

Reading icons, scores, and evidence

AI match insights are directional and qualitative.

• Positive signals—an upward icon in green indicates strong supporting evidence for that criterion
• Negative signals—a downward icon indicates missing or conflicting evidence that hurts the match, such as a required skill the resource doesn't have
• Neutral or partial signals—a two-way icon indicates that evidence is incomplete or indirect

Each section also shows a score out of 5 summarizing how strong that dimension is overall.

Evidence tags show which source a signal came from. Point at a tag to see the source detail behind it. Tags aren't links—they reveal their detail on hover rather than opening the source document.

How AI match score explainability works

Match score explainability follows the same approach an experienced resource manager would: first understand what the demand calls for, then evaluate whether the resource meets those expectations using all available evidence.

Understand what the demand requires

The system identifies the most important criteria for the demand by analyzing its description and its calibration parameters, such as required skills, titles, seniority, and industry signals. From these inputs, AI selects the criteria that matter most, so the evaluation stays focused rather than covering every minor detail.

This is why calibration matters so much. An uncalibrated demand produces no match score, so there's nothing to explain—and a poorly calibrated demand produces reasoning aimed at the wrong things.

Gather all available evidence about the resource

Once criteria are defined, the system reviews the available resource data to find relevant evidence, including:

• Resume and structured profile data
• Interview feedback, including AI-conducted interviews (shown as an AI Interview source)
• Experience the resource has recorded on the demand itself

That last source is specific to Resource Management. Demand experience already contributes to the match score, and it's included in the explanation as well—so the reasoning reflects the same evidence the score was calculated from, rather than a narrower view of the resource.

Evaluate whether each criterion is met

For each criterion, the system determines whether the evidence is positive, meaning the resource clearly meets the expectation; negative, meaning it clearly doesn't; or neutral, meaning the evidence is partial, indirect, or inconclusive.

The final explanation is built around these findings, showing where the resource aligns strongly with the demand, where they fall short and why, and which specific evidence supports each conclusion.

The system knows the existing overall, skill, and experience match scores, but it doesn't recompute them. It uses them to keep the tone and balance of the explanation consistent with the underlying scoring—so a high score generally yields mostly positive items.

In Resource Management, explanations stream in progressively: statements appear one at a time as they finish generating, rather than the panel waiting for the whole result. Explanations are cached per resource-and-demand pair, so a second view loads immediately.

FAQs

General

What is AI match score explainability?

It provides natural-language insight into why a resource is a strong or weak fit for a demand. It breaks the match score into two dimensions—Skill relevance and Experience relevance—and presents evidence-backed explanations for each, so you can quickly understand the reasoning behind a score.

Does this feature change how match scores are calculated?

No. The underlying match score calculation is unchanged. Explainability is a presentation layer that adds human-readable reasoning on top of the existing score. It doesn't alter, recompute, or replace the score.

Is the explainability model the same algorithm as the match score?

No. Explainability uses a separate AI model designed to generate natural-language insight. It takes the existing match score, calibration data, and resource profile information into account, but it operates independently from the scoring algorithm.

Where and how to view explanations

Where can I see match score explanations?

On the demand pipeline, point at a match score for a compact summary with a shorter insight under each section. On the resource profile, the Highlights panel shows the full breakdown, and the Matched Demands widget shows it for each matched demand.

Is this available to the resource themselves?

No. Explanations appear only on the manager-facing surfaces listed above. Although the person being matched is an employee rather than an external candidate, nothing in this feature is shown to them.

Why is the panel still loading?

Explanations stream in, so a brief loading state is normal while statements are still generating. Keep the panel open until they appear.

Data sources

What data sources does match score explainability use?

The explanation draws on the resource's profile data, resume, and any experience recorded on the demand itself. Each criteria line carries evidence tags, so you can trace where an insight came from.

Does interview feedback affect the match score itself?

No. Interview feedback doesn't change the underlying match score. It's used by explainability to generate richer, more complete reasoning. The score and the explanation layer are separate.

Can I click an evidence tag to open the source?

No. Evidence tags reveal their detail when you point at them, showing which source contributed to that line. They don't link through to the source document.

Understanding the evaluation

What are Skill relevance and Experience relevance?

They're the two evaluation dimensions. Skill relevance assesses how well the resource's skills, competencies, and knowledge align with what the demand requires, including functional readiness and transferability. Experience relevance evaluates how closely their delivery history, prior roles, and seniority align with the demand's expectations.

What do the icons and colors mean?

An upward icon in green means strong supporting evidence for that criterion. A downward icon means evidence is missing or conflicting. A two-way icon means evidence is incomplete or indirect.

A resource I expected to rank highly scored poorly. What should I do?

Read the criteria lines first—they usually show which requirement the resource fell short on. Where the reasoning is weighing the wrong things, recalibrate the demand and check the scores again.

Enablement and availability

Does my organization need to opt in?

The feature is enabled per tenant and group, and it's on by default. It's controlled separately from the Talent Acquisition equivalent, so having one doesn't mean you have the other. Contact your account representative to confirm your organization's status.

Why do I see the score without an explanation?

The most common reason is that the demand isn't calibrated, so no score exists to explain—the panel says so and offers a way to calibrate. Another cause is a resource profile too sparse to reason from. The panel names which applies.

What happens if the feature is turned off?

Views revert to the previous match score display, which shows the score without the breakdown. Nothing needs to be restored.

Interpreting results

How should I interpret an explanation that seems inconsistent with the overall score?

Explainability aligns its tone and balance with the underlying score, so a high score generally shows more positive signals and a lower one highlights more gaps. Where something looks unexpected, check the evidence behind the individual lines for context.

Does the explanation update when resource data changes?

Explanations are cached for performance. When key inputs change—the demand's calibration, or the resource's profile or resume—the cached explanation refreshes so later views reflect the update.

Is there a disclaimer I should know about?

Yes. The panel notes that insights are evaluated from available evidence and that you should use discretion when reviewing them. Treat the feature as decision support, alongside your own judgment.