38% of Candidates Are Walking Out — Here's What AI Hiring Gets Wrong

A new Greenhouse report found that 70% of candidates weren't told AI was being used in their hiring process — and 38% walked away entirely when they encountered it. That's not a technology problem. That's a design problem. In this episode, we look at what Chipotle and Unilever did differently. Chipotle's AI hiring assistant Ava Cado pushed application completion rates from 50% to 85%, cut time-to-hire from 12 days to 4, and doubled total applications — all by designing the AI experience around the candidate, not the recruiter. Unilever saw candidate satisfaction jump over 25% with a similar candidate-first approach. The difference comes down to three things: transparency about AI use, real responsiveness during the process, and actual closure at the end. Right now, 51% of candidates who complete an AI interview never hear back at all. Tools like OVI's Milo agent — which runs audio-only screening chats with built-in candidate communication, starting at $99/month on the Starter plan — make closure a default, not an exception. If you're tracking time-to-fill and cost-per-hire but ignoring candidate drop-off and ghosting rates, you're managing half the pipeline. The companies winning right now asked a different question: not 'how can AI speed us up?' but 'how can AI make this better for the candidate?'