Jack & Jill
HireAIScore rates Jack & Jill 49 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed September 25, 2026. Jack & Jill ranks 30 of 86 vendors rated in Candidate sourcing AI, against a category average of 46. That score is drawn from 36 evidence items across 7 rubric criteria for Jack & Jill — 31 with a cited source, 5 recording that nothing was located.
Score breakdown
Jack & Jill, criterion by criterion
Each rubric criterion is scored out of 100 on the cited evidence, then weighted into the total.
Weighted contribution shown to the right of each bar.
Article 11 Technical Documentation
Weight 20% · category avg 48
44
+8.8
Bias Audit Transparency
Weight 18% · category avg 40
62
+11.2
FRIA Support
Weight 15% · category avg 31
28
+4.2
Data Governance Disclosure
Weight 15% · category avg 53
48
+7.2
Human Oversight Design
Weight 12% · category avg 56
58
+7.0
Post-Market Monitoring
Weight 12% · category avg 40
48
+5.8
Customer Documentation
Weight 8% · category avg 57
56
+4.5
Total weighted score
48.6(rounded to 49 for display)
Category avg is the mean raw score on that criterion across the 86 Candidate sourcing AI vendors in scope of this rubric, this one included.
Strongest criterion
Bias Audit Transparency
62raw / 100
Contributes 11.2 to the total. 62 against a 40 category average.
Weakest criterion
FRIA Support
28raw / 100
Contributes 4.2 to the total. 28 against a 31 category average.
Cited evidence
The evidence behind the score
Every score is backed by at least one cited piece of evidence. Open a criterion to see what was found, when, and where.
Download diligence recordEvidence ledger
- Items
- 36
- Documentation
- 23
- Audit report
- 5
- Public statement
- 3
- Absence
- 5
- With a source URL
- 31 of 36
- Source hostnames
- 5
- Fewest items
- 2
- FRIA Support
These figures measure how thoroughly Jack & Jill was reviewed, not how Jack & Jill performed — an absence row, recording that nothing was located, is counted like any other item.
Article 11 Technical Documentation
4 items44
- DocumentationCaptured Sep 25, 2026
The CEO's blog (5 July 2026) describes Jill's per-role pipeline of LLM 'gates', each with chosen candidate fields, a tiered-criteria prompt, a structured output and a composite filter, run as one LLM call per candidate, with protected characteristics stripped first and prompts readable in the brief page.
https://www.jackandjill.ai/blog/ai-sourcing-is-broken-by-design
- DocumentationCaptured Sep 25, 2026
The docs say the hiring brief holds the evaluation criteria every candidate is scored on, signals and anti-signals, and Jill's sourcing strategy and ranking logic, with a version history that can be restored.
- DocumentationCaptured Sep 25, 2026
Warden's audit report gives the only third-party system description: inputs are job criteria and the candidate résumé, and the output is a match score.
https://trust.warden-ai.com/jackandjill/ai-candidate-matching/nyc
- AbsenceCaptured Sep 25, 2026
No model card, AI policy or principles page, validation study for the match percentage, EU AI Act technical documentation or ISO 42001 claim was found in the docs hub, legal pages, llms.txt, llms-full.txt or the blog posts in the sitemap, and /security, /trust, /ai-policy and /responsible-ai all return 404.
Bias Audit Transparency
6 items62
- Audit reportCaptured Sep 25, 2026
Warden AI's public dashboard for Jack & Jill's AI Candidate Matching shows one evaluation, dated 8 January 2026 (22,218 test samples, 15 protected categories plus sex by race, 102 Clear results), although it labels the audit frequency 'Monthly'.
https://trust.warden-ai.com/jackandjill/ai-candidate-matching
- Audit reportCaptured Sep 25, 2026
The NYC LL 144 report (8 January 2026) was run on Warden's own test profiles 'due to a lack of access to historical data' and shows sex impact ratios of 1.00 and 0.98, race ratios of 0.97 to 1.00, and a lowest intersectional ratio of 0.93 (Asian men).
https://trust.warden-ai.com/jackandjill/ai-candidate-matching/nyc
- Audit reportCaptured Sep 25, 2026
The EU AI Act bias report (same date, test data) adds counterfactual consistency testing (100% for women, 99.36% for men) and says it is not evidence of full EU AI Act compliance.
https://trust.warden-ai.com/jackandjill/ai-candidate-matching/eu-ai-act
- DocumentationCaptured Sep 25, 2026
Warden's Assured Directory lists Jack & Jill's AI Candidate Matching as assured from 17 December 2025.
- Public statementCaptured Sep 25, 2026
The site footer's 'Bias Report' link leads to a downloadable marketing guide ('The 2026 guide to bias in AI for recruitment') with no audit results, auditor or dates, so it is not credited as an audit.
- Public statementCaptured Sep 25, 2026
The CEO's July 2026 blog says the company runs 'regular audits on our debiasing' and publishes its methodology at /jill/bias, which is the marketing guide page, not the Warden audit.
https://www.jackandjill.ai/blog/ai-sourcing-is-broken-by-design
FRIA Support
2 items28
- AbsenceCaptured Sep 25, 2026
No FRIA template, EU AI Act instructions for use or deployer-obligation guidance appears in the 13-guide docs hub, the Platform Terms, the Data Sharing Agreement, the privacy policy, llms.txt or the blog.
- Audit reportCaptured Sep 25, 2026
Warden's EU AI Act bias report is the only EU-specific material; it gives deployers bias-test results and a generic list of provider and deployer duties, not an impact-assessment template.
https://trust.warden-ai.com/jackandjill/ai-candidate-matching/eu-ai-act
Data Governance Disclosure
8 items48
- DocumentationCaptured Sep 25, 2026
The privacy policy (dated 10 August 2026) relies on contract and legitimate interests for matching (consent only for special-category data), says Jack's call recordings and transcripts are never shared with employers, rules out biometric voiceprints and third-party model training, and keeps legally required data for at most two years after a deletion request.
- DocumentationCaptured Sep 25, 2026
The same policy says public professional profiles from LinkedIn and similar sources are collected under legitimate interests and scored by AI for the Sourcing feature, with a right to object but no described way of notifying those non-users.
- DocumentationCaptured Sep 25, 2026
The CEO's blog says all candidate context is stripped of protected characteristics before any matching gate runs, but no list of stripped fields is published.
https://www.jackandjill.ai/blog/ai-sourcing-is-broken-by-design
- DocumentationCaptured Sep 25, 2026
Warden's case study says Jack & Jill redacts candidate information before matching to reduce proxy bias.
https://www.warden-ai.com/case-studies/jack-jill-stress-tests-agentic-ai-for-bias-risk
- DocumentationCaptured Sep 25, 2026
The Data Sharing Agreement (effective 28 January 2026) makes Jack & Jill and each hirer independent controllers, incorporates the UK Addendum for restricted transfers, and lists the job-seeker data categories shared with hirers.
- DocumentationCaptured Sep 25, 2026
The Platform Terms name OpenAI, Anthropic, Google and ElevenLabs as model providers and allow AI models to be trained on customer information in anonymised, aggregated form.
- Public statementCaptured Sep 25, 2026
The app FAQ says a candidate's profile, professional details and contact information are 'automatically shared' with matching employers by default unless the candidate opts out, which contradicts the privacy policy's opt-in visibility and mutual-consent contact sharing.
- AbsenceCaptured Sep 25, 2026
No SOC 2, ISO 27001 or other security certification, trust center, subprocessor list or DPIA summary was found: trust.jackandjill.ai, security.jackandjill.ai and jackandjill.safebase.io do not resolve, trust.drata.com/jackandjill returns 404, and app.vanta.com/jackandjill/trust redirects to a login page.
Human Oversight Design
8 items58
- DocumentationCaptured Sep 25, 2026
Hiring teams define evaluation criteria with Jill, can open any candidate to see their score on every criterion alongside experience and education, and decide whether to shortlist or pass.
https://www.jackandjill.ai/docs/search-and-candidate-discovery
- DocumentationCaptured Sep 25, 2026
The getting-started guide says Jill explains why she picked each candidate and adds 'small flags or things to verify'.
- DocumentationCaptured Sep 25, 2026
The Platform Terms tell hirers not to rely on any Jill suggestion 'without prior human intervention, review and approval' by a qualified team member, and make the hirer responsible for every decision to engage a candidate.
- DocumentationCaptured Sep 25, 2026
The CEO's blog says early gates cut candidates on location, visa and function before later gates judge subtler traits, and that every gate prompt is readable in the brief page, but nothing documented lets the hiring team review the candidates the gates exclude.
https://www.jackandjill.ai/blog/ai-sourcing-is-broken-by-design
- DocumentationCaptured Sep 25, 2026
The docs say Jill shortlists the best candidates each day and that if a team consistently passes on a certain academic or professional background 'she'll deprioritise similar profiles'.
- DocumentationCaptured Sep 25, 2026
The privacy policy says Jack & Jill uses automated decision-making for matchmaking and invites anyone who wants a decision reviewed by a human to email legal@jackandjill.ai.
- DocumentationCaptured Sep 25, 2026
A 'How selective Jill should be' setting lets hirers move her bar 'from broader to only the clearest matches', and 'Reset search approach' rebuilds the search from the candidates the hirer said yes to.
- DocumentationCaptured Sep 25, 2026
Introductions are consent-based: the candidate must accept through Jack before any contact details are shared.
Post-Market Monitoring
4 items48
- DocumentationCaptured Sep 25, 2026
An incident.io status page that the site does not link tracks five components (the website plus Jill's and Jack's web apps and emails) and lists six incidents from 2 July to 21 September 2026 with timestamped updates but no postmortems.
- DocumentationCaptured Sep 25, 2026
A security.txt file names security@jackandjill.ai as the security contact, valid to 10 February 2027.
- Audit reportCaptured Sep 25, 2026
Warden's dashboard advertises continuous bias auditing and post-market monitoring on a monthly cycle, but the only evaluation it publishes is dated 8 January 2026.
https://trust.warden-ai.com/jackandjill/ai-candidate-matching
- AbsenceCaptured Sep 25, 2026
No public changelog, model or prompt update log, or AI-incident channel was found (/changelog, /updates and /release-notes return 404), and the privacy policy's claim of regular outcome audits has no published results beyond the January 2026 Warden audit.
Customer Documentation
4 items56
- DocumentationCaptured Sep 25, 2026
A public docs hub has 13 guides covering the hiring brief, search and candidate discovery, introductions, pipeline management, role configuration, team permissions and Ashby, custom-ATS and Slack integrations.
- DocumentationCaptured Sep 25, 2026
Hirer-facing Platform Terms (effective 1 July 2026) explain how Jill works, fees, the AI subcontractors used and the hirer's duty of human review before acting on Jill's suggestions.
- DocumentationCaptured Sep 25, 2026
A Data Sharing Agreement sets out controller roles, international transfer terms and the categories of job-seeker data shared with hirers.
- AbsenceCaptured Sep 25, 2026
No customer guidance on NYC LL 144 candidate notices, Illinois HB 3773 notices, Colorado disclosures or EU AI Act deployer duties was found in the docs hub, legal pages, FAQ, llms.txt or blog.
Editorial notes
Company overview
Jack & Jill is the trading name of Tinker Tailor Talent Limited, a London company incorporated in March 2025 by Matthew Wilson (CEO) and Saaras Mehan (CTO). It raised a $20M seed round led by Creandum in October 2025 and a $40M Series A led by Air Street Capital in September 2026. Jack, a free AI career agent, interviews candidates by voice and chat and keeps a structured profile of each; the company says more than 380,000 people use him. Jill, the employer-side agent scored here, turns a job description into a hiring brief with evaluation criteria. She then runs a per-role pipeline of LLM 'gates' over Jack's network, and optionally over public LinkedIn-type profiles. She scores each candidate on every criterion to produce an overall percentage match, shortlists daily, and makes an introduction only when the candidate agrees. Hirers pay a 10% success fee on first-year salary; US hirers can buy plans instead.
Regulatory exposure
Jill decides which candidates a hiring team ever sees, so every regime tracked here applies. Under the EU AI Act it is an Annex III point 4 high-risk system with Jack & Jill as provider. Those duties now apply from 2 December 2027. The company offers UK/Europe hirer terms but publishes no AI Act material beyond Warden's bias report. New York is a core market, so employers filling NYC roles need an LL 144 bias audit and candidate notice. Warden's LL 144 report is dated 8 January 2026, so it supports use only until January 2027 unless refreshed, and Jack & Jill offers no notice guidance. Illinois (HB 3773 notice since 1 January 2026) and Colorado (SB 26-189 disclosures from 1 January 2027) add disclosure duties. Warden's Colorado page still cites the repealed SB 24-205. California's FEHA automated-decision rules (since 1 October 2025) expect four-year record retention, but the privacy policy keeps legally required data for at most two years after a deletion request. Under UK GDPR, Jack & Jill is the controller. It reuses what Jack gathers for Jill's matching on contract and legitimate-interest grounds, not consent. It admits automated decision-making for matchmaking, offering human review by email, and it scores public profiles of non-users under legitimate interests.
Path to a higher score
Publish the monthly Warden evaluations the dashboard promises; only the 8 January 2026 run is public. Extend the audit beyond résumé-only test data to what Jill actually uses: Jack's conversation-derived profiles, learning from employer passes, and public-profile sourcing. Point the footer 'Bias Report' link at the audit rather than the marketing guide. Publish a model card or technical note showing how gate verdicts and per-criterion scores become the match percentage, with validation data and known limits. Add deployer guidance: NYC LL 144 notice templates, Illinois and Colorado disclosures, and EU AI Act instructions for use and FRIA inputs. Obtain and publish SOC 2 or ISO 27001 and a subprocessor list. Make the app FAQ's default-sharing answer match the opt-in privacy policy. Explain how non-users whose public profiles are scored are notified. Let humans review candidates the gates exclude, add guardrails on learned deprioritisation and keep decision audit logs. Link the status page and publish a changelog and postmortems.
Regulatory frame
What applies to candidate sourcing AI.
Ranking outputs are inside Annex III §4 when they influence hiring decisions. GDPR Article 22 (automated decision-making) and the GDPR rights around profiling apply independently for EU candidates.
Compare
Jack & Jill against its nearest-scoring peers.
The closest scores in Candidate sourcing AI, side by side.
- Side-by-sideJack & Jill vs Indeed50F
- Side-by-sideJack & Jill vs Stepstone49F
- Side-by-sideJack & Jill vs Lever49F
- Side-by-sideJack & Jill vs Oracle Recruiting49F
Ranking
Others rated in Candidate sourcing AI
Ranked 30 of 86 by weighted total under rubric v1.0. The ordering is arithmetic on the rubric and carries no view on which tool suits a given hiring process.
- #27Stepstone49F
- #28Lever49F
- #29Oracle Recruiting49F
- #30Jack & JillThis profile49F
- #31Avature48F
- #32X0PA AI48F
Conflicts of interest
No vendor pays for placement, scoring, or removal. Casework - the consulting firm that operates this directory - provides paid services to some vendors. Any active or recent (within 24 months) commercial relationship is disclosed on the affected vendor profile and the review is reassigned to an independent reviewer. See the full policy on About.
Casework has no commercial relationship with this vendor.