Explore the pilotBrief
Agree the vacancy, must-haves, useful experience and the evidence framework before screening begins.
The solution
Evidence between application
and interview.
InterventionAI combines technology for the volume with meaningful human review—turning applications into a focused, explainable shortlist employers can act on.
AI recommends. People decide.
1,057applications
Human
review
review
Evidencedecision-ready shortlist
The InterventionAI method
One managed evidence screen.
Five accountable steps.
The employer keeps control of the role criteria and final decision. InterventionAI handles the high-volume evidence work between them.
Structure
Organise applications consistently and surface role-relevant claims, gaps and duplication.
Review
A trained person assesses the evidence against the agreed framework and challenges automated findings.
Validate
Use short candidate conversations where key claims need testing or context needs clarifying.
Shortlist
Deliver evidence cards, clarification flags and interview prompts—not another pile of CVs.
Designed around the failure points
Every market concern needs a visible control.
The answer is not to remove AI from recruitment. It is to stop AI support from becoming an unchallengeable decision process.
Market concern
Strong candidates filtered out
InterventionAI response
Review every application against explicit essential and desirable criteria, with evidence considered at criterion level.
Market concern
Opaque automated rejection
InterventionAI response
Keep AI-proposed findings separate from the final human outcome. AI cannot silently write the progression decision.
Market concern
No meaningful human involvement
InterventionAI response
Require an authorised person to review every criterion before a final screening recommendation can be recorded.
Market concern
Human and AI disagree
InterventionAI response
Record whether the reviewer accepts, amends or rejects the AI finding, with a compulsory reason for every override.
Market concern
Decisions cannot be explained
InterventionAI response
Preserve the evidence, reviewer, date, outcome and material changes in an auditable record.
Market concern
CV polish distorts results
InterventionAI response
Ask applicants for structured evidence against the role criteria instead of relying on keyword matching alone.
Controlled POC statusThe client portal, application capture, criterion-level human review, override recording and audit trail are built. AI evidence extraction is the next controlled stage and will be tested with fictional applications before any live pilot use.
What the employer receives
An evidence card—not another CV pile.
Every recommendation is designed to make the interview sharper, not to make the final hiring decision.
- Demonstrated evidence against agreed criteria
- Clarification flags that do not hide uncertainty
- Targeted interview prompts for each candidate
- Human reviewer sign-off
Candidate 07
Senior Delivery Manager
- Eligibility
- Verified
- Stakeholder leadership
- Strong evidence
- Delivery recovery
- Evidence + one clarification
- Work scenario
- Meets framework
Interview prompt
What changed after your recovery plan?
Fairness by design
Trust comes from a consistent process—not an AI detector.
“We do not ask whether AI wrote the application. We ask whether the applicant can substantiate it.”
Same evidence framework
Applicants are considered against the same role-specific criteria.
Human accountability
A person reviews the evidence and owns every recommendation.
Proportionate data
Only information relevant to the screening decision should be collected.
Auditable handover
Employers see reasons, evidence and uncertainty—not a black-box score.
Prove it on one vacancy
Start with a controlled, human-reviewed pilot.
Agree the vacancy, evidence framework and application-volume band. Then measure the time saved and the quality of what reaches interview.
Explore the founding pilot