§ The result
Juicebox (PeopleGPT) leads Saville Assessment by 9 points.
Under rubric v1.0, Juicebox (PeopleGPT), filed under Sourcing, finishes on 57 and Saville Assessment, filed under Assessment, on 48, 9 points apart. Juicebox (PeopleGPT) leads 4 of the 7 criteria, including Bias Audit Transparency, while Saville Assessment leads 2, including FRIA Support, and 1 is level. Bias Audit Transparency is both the widest raw gap and the heaviest weighted swing — Juicebox (PeopleGPT) 84, Saville Assessment 42 — and at an 18% weight it carries 74% of the weighted advantage Juicebox (PeopleGPT) holds across the 4 criteria on that side of the table.
Vendor profile
Juicebox (PeopleGPT)
Juicebox (PeopleGPT) is an AI talent-sourcing search engine that lets recruiters find, rank, and reach candidates across 800M+ public profiles using natural-language queries instead of Boolean search, with an "Autopilot" agent that scores and ranks profiles against job criteria.
Vendor profile
Saville Assessment
Saville Assessment publishes the Wave personality questionnaires, Swift aptitude tests and situational judgment tests. They are scored automatically against norm groups and delivered through the Oasys platform, the Wave Connect hiring platform (which uses AI to build role profiles from job descriptions) and a two-minute Screener that returns a fit band into the employer's ATS.
§ The record
Juicebox (PeopleGPT) is filed under Candidate sourcing AI, Saville Assessment under Assessment AI. They also share 1 secondary category: Screening.
Both are in scope of EU AI Act, Colorado AI Act, NYC LL 144 and Illinois HB 3773. Only Saville Assessment is in scope of Mobley v. Workday.
| Record | Juicebox (PeopleGPT) | Saville Assessment |
|---|---|---|
| Price tier | Mid | Enterprise |
| Headquarters | US | GB |
| Founded | 2022 | 2004 |
§ What decides it
The criteria that move the gap, heaviest swing first.
18% weight · 66% of the movement so far
Bias Audit Transparency
Juicebox (PeopleGPT) 84 · Saville Assessment 42
7.6 weighted points to Juicebox (PeopleGPT)
12% weight · 83% of the movement so far
Post-Market Monitoring
Juicebox (PeopleGPT) 55 · Saville Assessment 38
2.0 weighted points to Juicebox (PeopleGPT)
15% weight · 90% of the movement so far
FRIA Support
Juicebox (PeopleGPT) 35 · Saville Assessment 40
0.8 weighted points to Saville Assessment
12% weight · 95% of the movement so far
Human Oversight Design
Juicebox (PeopleGPT) 50 · Saville Assessment 55
0.6 weighted points to Saville Assessment
20% weight · 99% of the movement so far
Article 11 Technical Documentation
Juicebox (PeopleGPT) 52 · Saville Assessment 50
0.4 weighted points to Juicebox (PeopleGPT)
15% weight · 100% of the movement so far
Data Governance Disclosure
Juicebox (PeopleGPT) 56 · Saville Assessment 55
0.1 weighted points to Juicebox (PeopleGPT)
Weighted delta is the raw difference times the criterion weight, so the 6 figures above sum to +8.8 — the difference between the two totals before either is rounded. Shares are of 11.5 points of movement counted without direction. The table below rounds each side to a whole number first, so its Δ raw total reads +9 beside that +8.8.
§ Category-by-category
Raw score per criterion, with weighted delta.
Δ is Juicebox (PeopleGPT) minus Saville Assessment.
| Category | Weight | Juicebox (PeopleGPT) | Saville Assessment | Δ raw | Δ weighted |
|---|---|---|---|---|---|
| Article 11 Technical Documentation | 20% | 52 | 50 | +2 | +0.4 |
| Bias Audit Transparency | 18% | 84 | 42 | +42 | +7.6 |
| FRIA Support | 15% | 35 | 40 | -5 | -0.8 |
| Data Governance Disclosure | 15% | 56 | 55 | +1 | +0.1 |
| Human Oversight Design | 12% | 50 | 55 | -5 | -0.6 |
| Post-Market Monitoring | 12% | 55 | 38 | +17 | +2.0 |
| Customer Documentation | 8% | 60 | 60 | - | - |
| Total | 100% | 57 (F) | 48 (F) | +9 | +8.8 |
§ 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 | Juicebox (PeopleGPT) | Saville Assessment | ||||
|---|---|---|---|---|---|---|---|
| Cited | Absences | Audit / hosting | Cited | Absences | Audit / hosting | ||
| Article 11 Technical Documentation | 20% | 1 | 1 | Sources on a separate domain | 5 | 1 | Audit reportSources on the vendor's own domainSources on a separate domain |
| Bias Audit Transparency | 18% | 3 | 0 | Audit reportSources on a separate domain | 4 | 1 | Audit reportSources on the vendor's own domainSources on a separate domain |
| FRIA Support | 15% | 1 | 1 | Sources on a separate domain | 3 | 1 | Sources on the vendor's own domain |
| Data Governance Disclosure | 15% | 2 | 0 | Sources on the vendor's own domain | 5 | 1 | Sources on the vendor's own domain |
| Human Oversight Design | 12% | 1 | 1 | Sources on a separate domain | 6 | 0 | Sources on the vendor's own domain |
| Post-Market Monitoring | 12% | 2 | 0 | Audit reportSources on a separate domain | 5 | 1 | Sources on the vendor's own domain |
| Customer Documentation | 8% | 2 | 1 | Sources on the vendor's own domain | 6 | 1 | Sources on the vendor's own domain |
| All criteria | 100% | 12 | 4 | Audit reportSources on the vendor's own domainSources on a separate domain | 34 | 6 | Audit reportSources on the vendor's own domainSources on a separate domain |
Juicebox (PeopleGPT) has 16 evidence rows across 7 of 7 criteria — 12 with a source URL, 4 documented absences; Saville Assessment has 40 evidence rows across 7 of 7 criteria — 34 with a source URL, 6 documented absences; some documentation or audit rows for Juicebox (PeopleGPT) and Saville Assessment sit on a domain separate from the vendor's own site.
Provenance
The scores above are documented by 56 evidence rows on file for these two vendors: 16 for Juicebox (PeopleGPT), across 4 distinct source domains, and 40 for Saville Assessment, across 2 distinct source domains.
- Rubric
- v1.0
- Juicebox (PeopleGPT) last reviewed
- June 7, 2026 · HireAIScore editorial team
- Saville Assessment last reviewed
- September 25, 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.