In the not-too-distant future, a hiring manager taps his watch and says, "Hal, have you hired the perfect candidate yet?" A cheerful voice replies, "Yes, Dave, I have found the right fit."

Plenty of job seekers think this isn’t the future—it’s already here. But is it really? To see where we are today, it helps to look back at how ATS tools originally worked: reacting to resumes as they came in, using “knock-out” questions and keyword searches to help recruiters prioritize candidates.

Now, with generative AI, hiring can be more proactive. AI agents can be set up before a job post even goes live, screening resumes as they roll in. This shift isn’t here to replace human judgment. Instead, it gives recruiters—especially solo recruiters or those juggling a mountain of requisitions—a helpful boost, handling those first rounds of screening so they can focus on the human side of the process.

“Hiring is like a box of chocolates—you never know what you’re gonna get.”

For years, companies have searched for ways to make hiring more predictable, hoping each new tool might be the answer. Each new hiring trend shows up with the promise of making those chocolates a little less random—whether it’s behavioral interviews, structured assessments, or pre-employment testing. But none of these have been silver bullets. Companies still end up using a mix of these tools based on their needs, size, and industry, hoping it’ll lead to quality hires.

With the arrival of Generative AI, there is a thought that this could be what finally changes the “box of chocolates” nature of hiring, bringing more predictability by pinpointing the “right fit” every time. Many believe it could. LLMs are built to handle large volumes of hiring data, from parsing resumes to analyzing candidate responses, quickly and efficiently. Advocates for AI in hiring think these models can uncover patterns that traditional methods might miss, allowing companies to select candidates with the ideal blend of skills and values for a role, potentially removing some of the guesswork from hiring decisions.

Going from improving speed and efficiency of the hiring process to hiring the best person for a role is a bold claim. Bringing someone new into the company is a mix of many things. Depending on the role and the level, an interview can take many shapes. Early-career candidates are often probed on "what" they can do, whereas senior-level candidates are more likely questioned on "why" they did it that way.

One of the more interesting parallels between behavioral interviewing and Generative AI lies in the art of asking the question or crafting the prompt. Both require thoughtful preparation in order to ask the right question or write a good prompt. In behavioral interviewing, recruiters craft specific questions to uncover how candidates have acted in past situations, using tools like the STAR (Situation, Task, Action, Result) method to guide them. The aim is to draw out relevant behaviors and insights that predict future performance. Miss the mark on the question, and you may miss the mark on the insight.

But this requires a degree of expertise. In behavioral interviewing, it’s the interviewer’s ability to ask the right questions that makes all the difference. Their skill in constructing and delivering questions directly impacts the quality of information they get from the candidate. Similarly, with generative AI, crafting the “right” prompt is key. The person using AI needs prompt fluency: the skill to design prompts that pull out meaningful responses. Without this expertise, whether in interviewing or prompting, the output can fall flat, lacking the depth and relevance needed to make informed hiring decisions.

And this is where the fear around AI comes in, especially for job seekers. There’s no transparency in how the decision gets made. The AI’s prompt box feels like a black box; what happens under the hood isn’t visible to the candidate. Relying only on AI for hiring decisions creates issues, as candidates are left in the dark about why they weren’t selected, with no real feedback or understanding of what factors the AI used to assess them.

While some argue that AI can make “nuanced” hiring decisions by analyzing complex data points, it’s crucial to ask whether this is realistic—or even desirable. True nuance in decision-making involves layers of context and insight that are often informed by human judgment.

Nuance doesn’t simply come from data volume; it’s about the ability to interpret and weigh subtle, often unquantifiable factors.

Even if you added more contextual data like a transcript from the interview itself, AI still lacks the ability to read between the lines, listen for tone in responses, or interpret the candidate’s body language.

The challenge that AI faces in finding the “right fit” for hiring calls to mind an episode of the TV show Elementary, where, to paraphrase Sherlock Holmes, “in large numbers, patterns start to emerge, while individuals remain unpredictable.” As Holmes points out, while we, like an LLM, might make educated guesses about what the “average person” would do, understanding one unique individual is far more elusive. This distinction captures AI’s core limitation in hiring: while it excels at recognizing broad patterns, it can overlook the unique qualities that make each candidate an individual, missing the nuance of true compatibility.

Any process that involves two or more humans is, by its nature, complex. And in hiring, that complexity only grows. Can companies improve how they hire? Absolutely. Does this shiny, new AI seem like it could help? You bet.

But using Generative AI to find the ‘right fit’ risks putting efficiency over empathy, and in hiring, I’m sorry, Hal, I’m afraid we can’t do that.