Eminence Growth Solutions, home

Search

Subscribe
Menu

Search

Subscribe

Software vendor

Harver

Harver sells hiring assessments that produce a Matching Score, and publishes a BABL AI bias audit summary dated November 4, 2025; not an SEC reporting company.

1 offering in the directory Last verified

Function

Ownership
Not stated

Not stated in the sources read. EDGAR returns no company named Harver.

Headquarters
Not stated

Not stated in the sources read.

SEC reporting
No
SourceSEC EDGAR company searches: Greenhouse Software, Harver, Outmatch (queries listed in the excerpt)
Audit firm
No
Website
harver.com
SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.
Subprocessor list
Not stated
Contract terms
Not stated

Offering · Software

Harver predictive assessments and matching

Directory recordLast verified Compare with othersBuild questions for this vendor

Harver's platform evaluates a candidate's personality, cognitive abilities and job-specific skills to produce a Matching Score and a Recommended outcome, with module weights and scoring thresholds that users can configure. SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.

Harver says it also uses AI for spoken language assessments, video interview analysis, explanations of the traits measured and automated steps such as stage progression and interview scheduling. SourceResponsible AI in Hiring

Harver publishes a bias audit summary by BABL AI, dated November 4, 2025, on April to May 2025 data that quantified impact by gender and race/ethnicity only. SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.

Its reported impact ratios range from 0.750 to 1.000, with groups under 2% of applicants not reported. SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.

Where it fits

CoE analysis Fits high-volume hiring where assessments and a match score sort applicants before interviews. SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.

CoE analysis The published audit used one recommendation threshold, the 46th percentile, and a two-month data window, so a buyer that sets different weights or cut-offs needs its own monitoring. SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.

What it does

Offering type
Software
SourceResponsible AI in Hiring
Workflow areas
  • Assessment and testing
  • Screening and ranking
  • Interviewing
SourcesSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.Responsible AI in Hiring
Customer size
Not stated

AI

Kind of AI
  • Established machine learning and rules
  • Built into the platform
SourceResponsible AI in Hiring
Suggests or acts
Suggests and acts
SourceResponsible AI in Hiring
Availability
Not stated

The sources read do not label features by availability or package.

Models named
Not stated

Control

Who controls it
  • Vendor
  • Customer HR team

Harver builds and validates the models. Its bias audit summary lists module weights and scoring thresholds as user-configurable settings that the customer sets; the customer role is not named.

SourcesSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.Responsible AI in Hiring

Commercial

Pricing model
Not stated
Public price
Not stated

Evidence

Independent outcomes
None found in the sources read.
Sources read
  • Vendor-stated 10
  • Independent market evidence None read
  • Independently verified outcomes None read
  • Pricing None read
  • Known limitations None read
Coverage
Directory record
Last verified

Decisions about people

Decision use
  • Ranks or scores
  • Recommends

A numerical Matching Score and a categorical recommendation; the audit computed selection rates on a Recommended outcome at or above the 46th percentile. Whether stage progression can follow the score without a person's review is recorded as a conflict.

SourcesSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.Responsible AI in Hiring
Bias audit published
Yes

Vendor-published evidence. No newer summary was found on the Harver pages read.

SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.
Traits the tool infers
Personality, cognitive ability and job-specific skills from assessments; Harver also lists video interview analysis and spoken language assessment among its AI uses.

Which traits video interview analysis evaluates, and from which inputs, is not stated.

SourcesSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.Responsible AI in Hiring
Configurable thresholds
Yes

Options and weights for assessment modules and scoring thresholds for recommendations are user-configurable settings that can affect output.

SourceSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.
Candidate notice support
Not stated

Not stated in the sources read; the bias audit summary says it does not cover candidate notification.

Accommodation route
Not stated

Not stated in the sources read.

Human review
Harver describes its AI as decision-support with human judgment central, and provides explanations of what traits are measured and why candidates receive recommendations.

A reviewer workflow for changing an adverse outcome is not described.

SourceResponsible AI in Hiring

Security and control

Single sign-on
Not covered in this record
User provisioning
Not covered in this record
Role-based access
Not covered in this record
Administrator controls for AI
Not covered in this record
Audit logs
Not covered in this record
Certifications
Not covered in this record
Data retention
Not covered in this record
Data residency
Not covered in this record
Training on customer data
Not covered in this record

AI features

As the vendor or firm describes them. Notes in small print record what the sources leave open.

  1. Matching Score (predictive assessments and matching)

    Combines personality, cognitive-ability and job-specific skills assessments into a Matching Score and a categorical recommendation for a role.

    Kind
    Established machine learning and rules
    Suggests or acts
    Suggests
    Who controls it
    Vendor, Customer HR team
    Availability
    Not stated

    Control: Harver builds and validates the models; module weights and scoring thresholds are user-configurable, and the customer role that sets them is not named.

    SourcesSummary of Bias Audit Results: Audit of Harver's Platform for New York City's Local Law 144, V1.0, 11/04/2025, prepared by BABL AI Inc.Responsible AI in Hiring

  2. AI assessments (spoken language, video interview analysis)

    Uses AI to evaluate candidates' suitability, including spoken language assessments, video interview analysis and models predicting job success from incumbent data.

    Kind
    Established machine learning and rules
    Suggests or acts
    Suggests
    Who controls it
    Vendor
    Availability
    Not stated

    Control: Not stated.

    SourceResponsible AI in Hiring

  3. Explanations of results

    Explains what traits are measured and why candidates receive specific recommendations.

    Kind
    Built into the platform
    Suggests or acts
    Suggests
    Who controls it
    Vendor
    Availability
    Not stated

    Control: Not stated.

    SourceResponsible AI in Hiring

  4. Automation (stage progression, scheduling, fraud detection, chatbot)

    Automates repetitive recruitment steps, including automated stage progression, interview scheduling, fraud detection and chatbot engagement with candidates.

    Kind
    Established machine learning and rules
    Suggests or acts
    Acts
    Who controls it
    Not stated
    Availability
    Not stated

    Control: Not stated.

    SourceResponsible AI in Hiring

Unclear or conflicting statements

Where the sources are unclear or disagree, both statements are shown with their links.

Whether candidates progress without a person's decision

Statement A

Responsible AI page: Harver's AI is decision-support, not decision-making, and human judgment remains central.

harver.com/why-harver/responsible-ai-in-hiring/
FAQ: What type of oversight is in place for AI decisions?

Statement B

Same page lists 'automated stage progression' and 'automatic interview scheduling' as AI uses, while its FAQ says scheduling and screening are rule-based automation, not AI (C-30468).

harver.com/why-harver/responsible-ai-in-hiring/
Harver's Use of AI: Automating; FAQ on automation and AI

Recorded under D-143, not resolved. Whether a candidate can move or stop at a stage on a Matching Score threshold without a person's review is not stated; Harver separates rule-based automation from AI (C-30468).

Unresolved · Both statements are published as found; CoE draws no inference about which governs.

Questions to ask

For this offering

  • Is there a bias audit newer than the November 4, 2025 summary, and does it test the weights, thresholds and jobs we will use rather than the 46th-percentile setting?
  • Can we export applicant, stage and outcome data by requisition to monitor adverse impact by gender, race/ethnicity, age and disability ourselves?
  • Which validation studies support each assessment module for our job families, and are they validated for screening out or only for ranking?
  • Which steps can progress or stop a candidate automatically on a Matching Score or knockout rule, who configures that, and is each change logged?
  • What does video interview analysis evaluate, from which inputs, and can it be switched off by jurisdiction?
  • How do candidates with a disability request an accessible format or an alternative to timed, game-based or spoken assessments?

From the diligence question set

  1. DQ-37 Has an independent auditor tested this tool for adverse impact, on data like ours? Can we see the full results: the groups tested and excluded, the data and its dates, the thresholds used, and the auditor's relationship to you?
  2. DQ-38 Can we monitor adverse impact on our own applicant flow, by job and by stage, and export the data to do it?
  3. DQ-39 What evidence shows that the tool measures something related to the job, and was it validated for the way we will use it: as a pass-or-fail screen, or to rank people?
  4. DQ-40 Which settings change who passes: weights, thresholds, knockout questions, cut scores? Who can change them, are changes logged, and does a change trigger a new audit?
  5. DQ-41 What notices, explanations and alternative processes does the tool support, and can a person with authority review and change an adverse decision?
  6. DQ-42 How can a candidate with a disability take the assessment in an accessible format or request an alternative, and does speed, voice, video or a timed task affect the score?
  7. DQ-43 Does any feature infer emotion, personality or other traits from face, voice or video, or use biometric data? Can each such feature be switched off?
  8. DQ-44 What will you commit to in the contract to support our obligations: audit cooperation, documentation of intended use and limits, notice support, records, and cooperation if a decision is challenged?

Build questions for this vendor