AI Screening Cuts Time-to-Hire by 43%—But Here's What the Data Actually Shows

The ROI case for AI screening has moved from vendor case studies to independent research—and the picture is more nuanced than the marketing suggests. In this episode, we break down what 2025 and 2026 enterprise studies actually show about where AI screening saves time, where it improves quality, and where the gains quietly disappear. Aptitude Research tracked 350 enterprise hiring programmes and found AI screening reduced time-to-shortlist by 43%. Lighthouse Research followed 180 mid-market employers and saw 23% off total time-to-hire. But the gains concentrate in early-funnel activities—CV triage and initial screening—while time-to-offer and time-to-accept barely move. AI is a volume tool, not a relationship accelerator. On quality-of-hire, a major SHRM meta-analysis found structured screening—whether human or AI—outperforms unstructured approaches by 18 to 25% on 90-day retention and early performance outcomes. The AI advantage isn't intelligence; it's consistency. That said, 12-month retention gains are less reliable, and two patterns reliably kill ROI in real deployments: running default rubrics (62% of enterprise tools never get customised), and candidate dropout rates of 30 to 66% in AI-only pipelines. Voice-based and conversational tools see 22 to 28% lower dropout than form or pre-recorded video alternatives—which is exactly why OVI's screening agent Milo runs live audio conversations rather than video or forms. The episode closes with four practical principles for getting real ROI from AI screening: customise rubrics before launch, monitor dropout by demographic, embed AI in a structured hiring process, and benchmark against your own baseline.