§ The result
Harver leads impress.ai by 18 points.
Harver finishes 18 points clear of impress.ai under rubric v1.0 — 56 against 38 — across the Assessment and Screening cohorts. Harver leads 5 of the 7 criteria, including Bias Audit Transparency, while impress.ai leads 1, including Human Oversight Design, and 1 is level. Bias Audit Transparency is both the widest raw gap and the heaviest weighted swing — Harver 80, impress.ai 22 — and at an 18% weight it carries 55% of the weighted advantage Harver holds across the 5 criteria on that side of the table.
Vendor profile
Harver
Pre-employment assessment suite covering cognitive ability, behavioral, and situational judgement, including the legacy Pymetrics games product.
Vendor profile
impress.ai
impress.ai is a Singapore-headquartered recruitment automation platform that uses conversational AI assistants, resume parsing and scoring, and structured competency-based screening to filter, rank and shortlist job candidates, integrating with enterprise ATS systems.
§ The record
Harver is filed under Assessment AI, impress.ai under Screening AI. They record no secondary category in common.
Both are in scope of NYC LL 144, Illinois HB 3773, EU AI Act, Colorado AI Act and Mobley v. Workday.
| Record | Harver | impress.ai |
|---|---|---|
| Price tier | Enterprise | High |
| Headquarters | NL | SG |
| Founded | 2015 | 2017 |
§ What decides it
The criteria that move the gap, heaviest swing first.
18% weight · 54% of the movement so far
Bias Audit Transparency
Harver 80 · impress.ai 22
10.4 weighted points to Harver
20% weight · 85% of the movement so far
Article 11 Technical Documentation
Harver 65 · impress.ai 35
6.0 weighted points to Harver
15% weight · 93% of the movement so far
FRIA Support
Harver 35 · impress.ai 25
1.5 weighted points to Harver
12% weight · 98% of the movement so far
Post-Market Monitoring
Harver 40 · impress.ai 32
1.0 weighted points to Harver
12% weight · 99% of the movement so far
Human Oversight Design
Harver 50 · impress.ai 52
0.2 weighted points to impress.ai
8% weight · 100% of the movement so far
Customer Documentation
Harver 60 · impress.ai 58
0.2 weighted points to Harver
Weighted delta is the raw difference times the criterion weight, so the 6 figures above sum to +18.8 — the difference between the two totals before either is rounded. Shares are of 19.3 points of movement counted without direction. The table below rounds each side to a whole number first, so its Δ raw total reads +18 beside that +18.8.
§ Category-by-category
Raw score per criterion, with weighted delta.
Δ is Harver minus impress.ai.
| Category | Weight | Harver | impress.ai | Δ raw | Δ weighted |
|---|---|---|---|---|---|
| Article 11 Technical Documentation | 20% | 65 | 35 | +30 | +6.0 |
| Bias Audit Transparency | 18% | 80 | 22 | +58 | +10.4 |
| FRIA Support | 15% | 35 | 25 | +10 | +1.5 |
| Data Governance Disclosure | 15% | 55 | 55 | - | - |
| Human Oversight Design | 12% | 50 | 52 | -2 | -0.2 |
| Post-Market Monitoring | 12% | 40 | 32 | +8 | +1.0 |
| Customer Documentation | 8% | 60 | 58 | +2 | +0.2 |
| Total | 100% | 56 (F) | 38 (F) | +18 | +18.8 |
§ Evidence
Cited sources and documented absences, criterion by criterion.
An absence is a finding, not a blank: the reviewer looked for a named artefact and recorded that it is not located. Where a source sits is a fact about hosting only.
| Criterion | Weight | Harver | impress.ai | ||||
|---|---|---|---|---|---|---|---|
| Cited | Absences | Audit / hosting | Cited | Absences | Audit / hosting | ||
| Article 11 Technical Documentation | 20% | 3 | 0 | Sources on the vendor's own domain | 2 | 1 | Sources on the vendor's own domain |
| Bias Audit Transparency | 18% | 2 | 0 | Audit reportSources on the vendor's own domain | 2 | 1 | Sources on the vendor's own domain |
| FRIA Support | 15% | 1 | 0 | Sources on the vendor's own domain | 2 | 1 | Sources on the vendor's own domain |
| Data Governance Disclosure | 15% | 1 | 0 | Sources on the vendor's own domain | 2 | 1 | Sources on the vendor's own domain |
| Human Oversight Design | 12% | 1 | 0 | Sources on the vendor's own domain | 3 | 1 | Sources on the vendor's own domain |
| Post-Market Monitoring | 12% | 1 | 0 | — | 3 | 1 | Sources on the vendor's own domain |
| Customer Documentation | 8% | 1 | 0 | Sources on the vendor's own domain | 4 | 1 | Sources on the vendor's own domain |
| All criteria | 100% | 10 | 0 | Audit reportSources on the vendor's own domain | 18 | 7 | Sources on the vendor's own domain |
Harver has 10 evidence rows across 7 of 7 criteria — 10 with a source URL, 0 documented absences; impress.ai has 25 evidence rows across 7 of 7 criteria — 18 with a source URL, 7 documented absences.
Provenance
The scores above are documented by 35 evidence rows on file for these two vendors: 10 for Harver, across 1 distinct source domain, and 25 for impress.ai, across 1 distinct source domain.
- Rubric
- v1.0
- Harver last reviewed
- May 21, 2026 · HireAIScore editorial team
- impress.ai last reviewed
- August 2, 2026 · HireAIScore editorial team
HireAIScore scores are editorial research, not legal advice, and no vendor pays for placement, scoring or removal — see the methodology, conflicts policy and terms.
§ Where next