What an AI Screener Taught Me About Building Products

I had just completed a rather draining AI recruiter screening. The success message: an invitation to “strengthen [my] candidate package” by sending a reference link to professional contacts who can “highlight [my] skills and experience.”

I still didn’t know the name of the hiring company.

How I got here was simple: I’d responded to a job posting for a nonprofit product leadership role. The recruiter turned out to be an “agentic recruiting firm” and not the human I’d assumed. Once I realized that, I considered bailing. But having served on a nonprofit board I understood the need for cost-efficient hiring. Besides, I was curious how it would go.

The first question of my screening was broad, open-ended, and contained four subquestions. The tool’s follow-up asked for information that I’d already provided, explicitly, in the last sentence of my original response.

Five more questions followed the same pattern. Then I was asked to invite my professional contacts into the same black box, with no indication a human would ever see any of it.

To be clear: The problem isn’t that AI was involved. The hiring company has real cost constraints. The people who designed the system are no doubt trying to identify the strongest candidates for the role.

The candidate, though, was never a stakeholder. Just an input.

Originally published on LinkedIn