Hiring: Trials Replace Resumes

The résumé is losing its old job. In parts of the startup world, employers are replacing polished bullet points with weeklong tryouts that put candidates in the office and ask them to do the work before they are hired. Business Insider reported on April 7 that some companies are using in-person trials, simulations, and referrals instead of leaning on application portals stuffed with AI-optimized résumés, a shift driven by a simple complaint from managers: they no longer trust what a CV says about what someone can actually do (businessinsider.com). That mistrust did not appear out of nowhere. The hiring funnel has been warped by automation on both sides. Employers use software to screen for keywords. Applicants use AI to generate those keywords at scale. The result is a market flooded with applications that look tailored and competent, even when they reveal very little about real ability. Business Insider described this as a “show your work” era of hiring, where companies increasingly want candidates to demonstrate judgment, collaboration, and output in conditions that feel like the actual job (businessinsider.com, inc.com). That logic is not new. Industrial-organizational psychologists have been saying for decades that direct assessments beat vague proxies. A classic 1998 meta-analysis by Frank Schmidt and John Hunter found that work-sample tests were among the strongest predictors of job performance, far ahead of education and years of experience alone. More recent work has refined the ranking, but not the core lesson: structured, job-relevant assessments predict performance better than unstructured interviews and résumé gloss (emilkirkegaard.dk, cambridge.org, siop.org). What is new is where this is happening. AI has turned ordinary knowledge work into a credibility problem. Companies say they need AI talent, but many still do not know how to test for it. A Just Badge research release published April 7 claims there are 3.2 AI roles open per candidate, yet only 2.2% of interviews actually assess the relevant skills. The firm argues that employers are screening for the wrong things, then declaring a talent shortage when they cannot find what they never measured (globenewswire.com, finance.yahoo.com). The specific numbers in that release come from a company with something to sell, so they should not be treated as neutral labor-market truth. But the broader pattern is easy to verify. Job postings mentioning AI have kept rising even as overall hiring has stayed weak, and employers are concentrating scarce openings around AI-linked skills. Indeed Hiring Lab reported in January 2026 that jobs mentioning AI were growing amid broader hiring softness. Lightcast has also documented fast growth in generative-AI-related demand across the labor market, not just in pure tech roles (hiringlab.org, lightcast.io). That makes the trial-hire trend easier to understand. Employers are not abandoning credentials because they became enlightened. They are doing it because the old signals got noisy. A résumé can now be mass-produced. A conversational interview can be rehearsed with a chatbot. A candidate sitting in the office for a week, shipping work under observation, is harder to fake. That is why the hiring process is starting to look less like paperwork and more like an audition (businessinsider.com, apa.org). The catch is that this method shifts risk onto applicants. A weeklong trial can reveal competence, but it also asks people to give time, travel, and often unpaid labor before they have an offer. That tradeoff is easiest for well-funded startups to ignore and hardest for candidates with jobs, caregiving duties, or no cushion. So the same system that promises a cleaner read on skill can also narrow who is able to compete. The proof-of-work economy may be more honest than résumé theater, but it is still an economy, and it still sends the bill somewhere (businessinsider.com, apa.org).