Why YoHe doesn't screen candidates with AI (yet)
It's the single most common question we get, and the honest answer is more useful than a roadmap slide: we haven't built it, and we're not going to pretend we have.
Every hiring tool in this category is racing to bolt on AI candidate scoring, and we get asked about ours constantly. The honest answer is that it doesn't exist yet, and we'd rather say that plainly than dress up a keyword filter as intelligence.
That's not a values statement about AI in hiring generally — it's a statement about what's actually running in production today. Today, YoHe reads nothing on your behalf. It stores candidates, tracks them through a pipeline, and gives you a career page. A person reads every application.
Why we didn't fake it for the launch
It would have been easy to ship a scoring bar next to each candidate — a percentage that goes up when a resume contains the right keywords. It would have demoed well. It also would have been close to meaningless, and once a number is on the screen, people start trusting it more than they should.
A fake signal is worse than no signal, because it changes behavior. If we're going to put a score in front of someone making a hiring decision, it needs to hold up when a candidate — or a regulator — asks what it was based on.
What we'd actually need to get right first
- A model can't infer protected characteristics from a resume, even by accident, and we need to be able to prove that, not just claim it.
- Every output needs to be labeled as AI-generated and stay fully editable — nothing quietly overrides a human's judgment.
- A recommendation is not a rejection. Nothing gets automatically declined without a person acting.
If and when AI screening ships, it'll be documented in the product's guardrails page before it ships, not after — and it'll be off by default.