The average cost of a mis-hire — accounting for recruitment, onboarding, lost productivity, and the cost of replacing the person — is estimated at 30–50% of first-year salary. For an AI role paying $90,000, that's $27,000–$45,000 per failed hire. The core cause is almost always the same: the skills on the CV didn't reflect the skills the candidate actually had.
The CV problem
CVs are self-reported. "Proficient in AI automation tools" means nothing without a demonstration. A candidate who spent three hours watching a Make tutorial and a candidate who has deployed 15 production workflows both describe themselves as proficient. The hiring manager reading the CV has no way to tell them apart — until the 3-month mark when one of them is still asking basic questions.
Verified, graded evidence changes this calculus. A project submission with a 91% score and a Distinction grade means something concrete: the candidate understood a real-world business problem, designed a workflow to solve it, selected appropriate tools, and documented the solution well enough that an AI grader could evaluate its completeness. That's not self-reported. That's demonstrated.
Three ways companies restructure their funnel
Company A (Marketing agency, 45 employees): Previously spent two rounds of interviews establishing basic AI literacy. Now uses Readiness Score as a first-pass filter. Candidates below 70 go to a standard interview process. Candidates above 80 skip straight to a culture and role-fit conversation, with the project submission used as a talking point. Time-to-hire dropped from 6 weeks to 3.5 weeks for AI roles.
Company B (Financial services, 200 employees): Uses sector certifications as a hard filter. Any candidate interviewing for a Finance AI role must have the Finance sector certification. This ensures every candidate at the interview stage has already demonstrated they can apply AI in a financial services context — not just in general. Rejection rates at the interview stage dropped significantly because candidates arriving are already pre-qualified.
Company C (Legal tech startup, 12 employees): Uses the live assessment room during the interview itself. Rather than hypothetical questions ("how would you build a contract review workflow?"), they generate a real project brief for the candidate's skill level and give them 45 minutes. The AI assessment report from that session becomes part of the hiring decision alongside the conversation. They describe it as "the closest thing to a working interview without asking someone to do free work."
The interview as confirmation, not evaluation
The common thread across all three companies is a mindset shift: the interview becomes confirmation of what the portfolio already shows, not the primary evaluation. This is both more accurate and more respectful of the candidate's time — they've already proved themselves through the portfolio. The interview is about fit, context, and conversation, not about trying to assess skills in 45 minutes under pressure.
Setting up your hiring funnel
If you're using the platform as an employer:
- **Set a minimum Readiness Score** as your first filter — 70 for junior roles, 80+ for mid-senior
- **Require at least one graded project** in your industry sector before advancing to a first call
- **Use sector certifications** as a hard filter if you need proven cross-sector competence
- **Run a live assessment** for final-stage candidates — use the AI report as a structured conversation starter, not a pass/fail
The goal isn't to automate hiring. It's to make the human conversations that matter count for more.