Lever (since 2022 part of Employ, which also owns JazzHR and Jobvite) positions itself differently than classical ATSes: Talent Relationship Management. The CRM-style approach — candidate nurture, talent pools, automated campaigns — got an AI layer when Employ rolled out AI features in sourcing recommendations, automated nurture and predictive analytics. For mid-market Dutch employers that means: the classification question has shifted from "is this an ATS issue?" to "which AI feature affects our candidate flow?".
What Lever publicly offers
Based on Lever and Employ product pages and release notes:
- Talent Relationship Management — CRM for candidates with nurture campaigns and pipelines
- Recommended candidates — AI suggestions for matching candidates on a requisition
- Automated nurture — AI-determined timing and content for candidate engagement
- Sourcing AI — external candidate discovery via integrated tools and chrome extension
- Predictive analytics — pipeline forecast, time-to-fill predictions
- AI-assisted candidate communication — generated messages for outreach and follow-up
Lever's philosophy: candidates are an ongoing relationship, not a one-off transaction. AI reinforces that relationship but must not replace human choice.
The seven checks applied to Lever
1. Does the AI rank or score candidates?
Recommended candidates produces rankings against open requisitions. For active deployers: 4(a).
2. Does the AI optimize who sees a vacancy?
Sourcing AI and automated nurture influence which candidates are targeted. Targeting within 4(a) scope.
3. Is CV parsing really only parsing?
Document parsing with field extraction and (increasingly) skills inference for matching. Check release version.
4. Is the chatbot logistical or selective?
Lever Conversational AI for automated nurture is a borderline case: a nurture bot that qualifies candidates for pipeline progression hits 4(a).
5. Does the assessment tool measure behavior or performance?
Lever itself does no psychometric assessments. Integrations with external assessment vendors have their own classification.
6. Does the system continue post-hire?
No, Lever is pre-hire. Employ's broader portfolio (JazzHR for SMB) touches other markets.
7. Can you substantiate the vendor claim?
Lever and Employ publish AI positioning but no Model Cards at enterprise level. Request written documentation via Customer Success.
The classification call
Lever deployments typically run:
- Lever as basic ATS without Recommended Candidates and without Automated Nurture AI: workflow and CRM, limited AI Act relevance
- Lever with active AI recommendations, automated nurture and sourcing AI: defensive 4(a) high-risk
For Dutch mid-market employers (250-2000 employees) using Lever as primary recruitment platform: a feature audit is needed and a classification decision to document.
Vendor due diligence for Lever
Do a Lever feature audit
CRM-style recruitment enables many features that are not always actively used. Start with an inventory of what really feeds recruiter decisions.
Split CRM workflow from AI decisioning
Talent Relationship Management workflow is largely not an AI Act issue. AI-driven recommendations and nurture are. Document them separately.
Build oversight around AI recommendations
Recommendations work as oversight only if recruiters genuinely read the output critically and override. Train on that, log overrides, and use the HR AI Evidence Pack for dossier.
Test your AI against the Article 6(3) filter
Interactive self-assessment, updated for the Commission guidelines of 19 May 2026. 9 steps, personal report with reasoning, vendor questions and next steps.
Frequently asked questions about Lever and the AI Act
Practical questions for mid-market employers with Lever under the EU AI Act.
What to do now
For Lever users in NL and EU mid-market: feature audit this week, written vendor confirmation within 30 days, oversight documentation via HR AI hub. The CRM-style of Lever is an advantage for your oversight — you already have structured processes to anchor your AI classification on.