As AI Standardizes Applications, Hiring Turns to Evaluating How Candidates Think

AI is creating a hiring dilemma that is compounding on both sides of the table. Entry-level roles that young professionals traditionally use to build experience are shrinking. The result is fewer opportunities for recent grads, rising Gen Z unemployment, and a flood of stressed applicants overwhelming hiring funnels. The imbalance is forcing companies to rethink how they hire today. But the longer-term question remains: if the bottom of the corporate ladder disappears, what does the candidate pool look like a decade from now?

It’s a problem Abhra Roy thinks about a lot, and one he believes will take employers, applicants, and universities working together to solve. Roy is an associate professor of economics at Kennesaw State University who specializes in non-cooperative game theory and mechanism design, and has spent more than two decades thinking critically about students entering the workforce. He also recently founded workforce intelligence firm EverGain, where he serves as founding president and CEO. Between his work at EverGain and on campus, he is making the case that today’s hiring challenges are part of a wider coordination problem between employers, workers, and universities in an economy where AI is spreading quickly and employers must rethink how they find and evaluate candidates.

“Everyone is going to have polished resumes. But if you can show in your work ‘here’s a problem and here’s how I think about it,’ that’s what makes a candidate stand out,” Roy shares. But that distinction is becoming harder to recognize.

AI’s prevalent use by applicants is making the resume less useful to HR teams. Thanks to LLMs, more candidates can generate near-perfect cover letters that also look practically identical. This homogeneity, the larger flow of applicants, and keyword-optimized AI-generated materials are breaking down the traditional keyword filter system that allows hiring managers to tell candidates apart.

• Keeping up with the algorithms: Recruiters are losing clear hiring signals exactly when they need it most, leaving many organizations scrambling to fix a broken hiring funnel just as they lean into AI to keep pace with a shifting global work landscape. The pressure to move fast is causing problems. “Organizations are adopting AI without thinking much ahead,” Roy says. “Because what they’re thinking is, ‘oh, everyone is adopting AI, and if I don’t, I’m going to be left behind.'”

For some organizations, the rush to keep pace with peers is leading to decisions about automation that do not fully account for longer term effects on skills and experience. That reactive automation is exactly why Roy suggests a more intentional framework of shifting attention away from what materials candidates submit and toward how they think in order to meet the new demands of the workforce.

• Stakes are high: Eliminating too many jobs at the entry and mid-level risks weakening the pipeline that feeds senior roles later. “If current patterns continue, firms could discover that all of their senior talent is about to retire, and they have nothing to replace it with,” Roy explains.

Roy suggests hiring managers implement updated evaluation systems that assess a candidate’s “shape of thinking,” or the reasoning process they use to navigate uncertainty. That moves the focus away from standardized resumes and toward visible problem-solving. His solution isn’t to ban AI, as it will continue to gain relevance for just about every role. Rather, he suggests HR teams start testing how people use it. He gives the example of giving applicants a hypothetical business problem and asking them to submit a short response along with the exact prompt they used to solve it. Roy developed what he calls a “prompt sophistication index” that scores the quality and structure of a candidate’s prompt, allowing employers to set threshold levels and filter large applicant pools accordingly.

• Prompt over polish: “Firms should start looking more at how someone thinks, because that is the only thing that the AI will not take,” Roy says. “How is someone processing an uncertain environment? That is going to be the critical thing for every firm going forward.”

• Show the math: Even without a formal scoring system, the prompt itself can be revealing. “If the prompt is very thin and the output is really strong, then it’s the AI doing all of the work,” he says.

Roy notes universities also play a critical role in solving for this AI problem. He says traditional tests do not accurately capture what students are learning anymore. If students use AI to bypass early learning stages, some may not develop the baseline expertise they need. Rather than framing the trend as a crisis of cheating, Roy views it as a tactical measurement problem for universities. He believes schools have a responsibility to ready students for the workforce by teaching responsible AI use and building systems that show whether AI is truly augmenting student work or simply masking gaps.

• Crutch or catalyst: “We need a better measurement system for what is happening to students’ learning when they use AI. Where exactly their understanding is being augmented, where it is being masked, or where they are using AI as a crutch,” Roy says.

For employers, the most practical response to AI in hiring is not to out-filter automation, but to make the human thinking behind it more visible. Clearer ways to see how people approach problems when generative tools are part of the process can replace the signals lost to polished cover letters. “The best thing to do going forward in the future is to understand the shape of thinking,” Roy says. “It’s far more important than a resume.”