Why Engineering Hiring Takes 89 Days — And How AI Audio Screening Fixes It

Engineering hiring is broken — and the phone screen is exactly where it breaks. The average engineering role takes nearly two months to fill. For AI and machine learning positions, that number climbs to 89 days. And the bottleneck isn't sourcing or final-round panels — it's that 30-minute manual phone screen happening hundreds of times per open role. In this episode, we look at how Greenhouse and Lever — the two ATS platforms engineering teams reach for first — handle (or don't handle) top-of-funnel screening. Both are capable systems. Neither has built native AI screening that replaces the phone call. So every candidate still requires a human, a calendar block, and a 24-48 hour feedback loop. Then there's OVI. Instead of managing the phone screen, OVI's Milo agent replaces it entirely. Candidates complete an asynchronous AI audio-only screening chat on their own schedule — no video, no scheduling, no engineer time spent until after the pool has already been filtered. For a team receiving 200 applications for a senior backend role, that changes the math dramatically. We break down the cost comparison (OVI starts at $99/month vs. enterprise pricing for Greenhouse and Lever), walk through how Milo's screening actually works, and explain why a five-minute AI audio chat at roughly $2.50 per candidate is a fundamentally different model from blocking an engineer's hour at $75–150 loaded cost. Plus: what OVI doesn't do (coding assessments), how it aligns with GDPR and automated employment decision regulations, and the three-step setup that has engineering teams screening candidates before a single calendar invite goes out.