Does GoVerse integrate with SmartRecruiters?

Yes. GoVerse integrates with SmartRecruiters with full read/write access: candidate profiles, CVs, stages, scores, notes, and rejection reasons. It imports both historical and new candidates, not just new applicants.

Can recruiters override AI scores in GoVerse?

Yes. Any AI score can be overridden by a recruiter at any time. The override is logged in the audit trail but the human decision always takes precedence. AI provides recommendations, humans make final decisions.

Can GoVerse explain why one candidate scored higher than another?

Yes. Every score includes a per-criterion breakdown, written reasoning in plain English, and a comparison view showing exactly why two candidates scored differently. This is the default, not an add-on.

Where is candidate data stored in GoVerse?

For Australian customers, all candidate data is processed and stored in AWS Sydney (ap-southeast-2). GoVerse is compliant with the Australian Privacy Act including APP 8 cross-border disclosure controls. Candidate data is never used for AI model training.

Does GoVerse make automated hiring decisions?

No. GoVerse makes recommendations (it scores and ranks candidates) but humans make all selection decisions. No candidate is automatically rejected by AI without human review. This aligns with Australian government guidance on AI in employment.

How does GoVerse test for AI bias in hiring?

GoVerse monitors score distributions across protected characteristics in real-time, applies EEOC four-fifths rule testing for adverse impact detection, validates criteria for discriminatory signals before screening runs, and alerts compliance admins when bias thresholds are crossed.

How GoVerse works

The answers procurement, compliance, and IT teams need before they sign off.

ATS Integration

GoVerse connects to your ATS via direct integration or bulk import. Here's what the integration covers:

Capability	Support
Read candidates	✓ Full profile, CV, attachments
Write scores back	✓ AI scores, notes, and reasoning pushed to ATS
Candidate stages	✓ Read current stage, trigger stage transitions
Rejection reasons	✓ Structured rejection with AI reasoning
Notes & activity	✓ AI screening results written as structured notes
Historical candidates	✓ Full historical import, not limited to new applicants
Bulk import	✓ CSV/bulk upload for any system without direct integration

Supported ATS: SmartRecruiters, Greenhouse, Workday, Lever, JobAdder, Bullhorn, BambooHR, and any system via CSV export.

How AI scoring is configured

Recruiters have full control over what the AI evaluates and how much weight each criterion carries:

•Recruiter-defined criteria: you specify what matters: skills, experience, qualifications, certifications. The AI doesn't decide what's important. You do.
•Weighted scoring: each criterion has a configurable weight (must-have, preferred, nice-to-have). Adjust anytime.
•Recruiter override: any AI score can be overridden by a human. The override is logged but the human decision takes precedence.
•Per-requisition configuration: different roles can have completely different scoring criteria. No one-size-fits-all model.
•Multi-model architecture: scoring uses multiple AI models to reduce single-model bias. If one model underperforms, others compensate.

Decision auditability

"Can you explain why Candidate A scored 82 and Candidate B scored 61?" Yes. Always.

Every score includes:

✓ Per-criterion breakdown: which criteria matched, how strongly, and what's missing
✓ Written reasoning: plain English explanation of the decision (not just numbers)
✓ Comparison view: side-by-side explanation of why two candidates scored differently
✓ Immutable audit log: timestamped, versioned, tamper-proof record of every decision
✓ Model version tracking: which AI model produced the result, allowing retrospective analysis
✓ Export: full decision history exportable for regulatory review

This is not optional or an add-on. Explainability is the default for every screening run.

Bias testing & adverse impact monitoring

GoVerse actively monitors for disparate impact across protected characteristics:

✓ Real-time statistical monitoring: score distributions analysed across gender, age, ethnicity, disability, and other protected groups
✓ Four-fifths rule testing: automatic detection of adverse impact using EEOC four-fifths threshold
✓ Pre-screening policy validation: criteria are checked for discriminatory signals before they're applied to candidates
✓ Alerting: compliance admins notified when bias thresholds are crossed
✓ Corrective recommendations: system suggests criteria modifications to reduce disparate impact

See our compliance page for full details on fairness controls.

Data processing & sovereignty

For Australian customers:

✓ Data stored in Australia: candidate data processed and stored in AWS Sydney (ap-southeast-2)
✓ Australian Privacy Act compliant: full APPs compliance including APP 8 (cross-border disclosure controls)
✓ No data used for model training: candidate data is never used to train AI models. Processing only.
✓ Data retention controls: configurable retention periods per account. Delete on request.
✓ Encryption: at rest (AES-256) and in transit (TLS 1.2+)
✓ Access controls: role-based access, audit logging on all data access

AI recommendations vs automated decisions

GoVerse makes recommendations. Humans make decisions.

This is a deliberate design choice aligned with Australian government guidance on AI in employment:

✓ AI scores and ranks: but does not automatically reject, progress, or hire anyone
✓ Recruiter reviews shortlist: the scored results are a starting point, not a final decision
✓ Override at any point: recruiters can override scores, re-order candidates, and make exceptions
✓ No automated selection: candidates are never automatically rejected by AI without human review
✓ Transparency to candidates: organisations can disclose AI use in their process (we recommend this)

Australian guidance alignment: The APSC AI principles and APS employment guidance require that AI supports rather than replaces human decision-making in selection. GoVerse is designed for this model: AI assists, humans decide.

The workflow, start to finish

The sections above answer the sign-off questions. Here is what actually happens day to day, from an empty requisition to a shortlist a hiring manager can act on.

Bring in the candidates

Connect your ATS (Workday, SAP SuccessFactors, SmartRecruiters, Greenhouse, Lever and others) and pull both live and historical candidates, or bulk upload CVs by CSV. Nothing is capped at "new applicants only", so an old pipeline becomes searchable again the moment it lands.

Write the criteria in plain language

Describe the role the way you would to a colleague: the must-haves, the preferred experience, the qualifications that matter. You set the weights (must-have, preferred, nice-to-have). The AI does not decide what is important, you do, and you can adjust it per requisition.

The policy engine checks before it runs

Before a single candidate is scored, the bias-aware policy engine reviews your criteria for discriminatory signals and flags anything that risks disparate impact. This happens up front, not after the fact, so a problematic filter is caught before it touches applicants.

Every candidate is scored and explained

The AI evaluates the full pool (not the first 20) and returns a 0 to 100 score for each person with a per-criterion breakdown and written reasoning. At roughly $0.01 per CV in credit terms, screening 200 applicants costs about two dollars, so there is no economic excuse for skimming.

Review, compare, and override

Start from the ranked list, open the reasoning on the strong candidates, and use the side-by-side comparison to see why one scored above another. Disagree with a score? Override it. The override is logged in the audit trail, but the human decision takes precedence.

Sync back and engage

Scores, notes, and reasoning are written back to your ATS as structured records, and stage transitions can be triggered from GoVerse. From there you move shortlisted candidates into engagement, so the strong ones hear from you before a competitor gets there.

Setup to first scored run is minutes, not a project. There is no implementation team and no field-mapping exercise. The recruiter who knows the role is the one who configures it.

Have more questions? Our team can walk through a technical deep-dive with your procurement and compliance stakeholders.

Talk to our team →Or try it yourself, free