If you’re a TA professional in 2025, there’s a good chance your team uses AI. Which probably feels awesome, right? Less time spent on boring administrative work. Less budget spent on job advertising. And after years of fighting a seemingly unwinnable battle in the war for talent, you’re finally starting to turn the tides. All of this is great!
Here’s the not so great part: Candidates have the same secret weapon.
Yes, more candidates than ever are using AI, too. In fact, a recent study by Capterra surveying 3000 applicants across 12 countries found that 58% of candidates used AI to help them land a job. And while most might be generating resumes and cover letters, some candidates are using AI for more, let’s just say, creative purposes (like faking job interviews).
So unfortunately some of that progress you’re making is being countered. It’s not necessarily that candidates are fighting dirty — but they are changing the rules. In a world where everyone can gain an upper hand with AI, does anyone actually win? Where do we go from here?
Well, I actually think we move forward by taking a step back.
Hiring is a people game. Now I’m a psychologist, so to me everything is a people game. But I really believe that making a great hire has never been about finding the right resume; it’s about finding the right person. AI isn’t the best at that, mostly because that’s not what it was built to do. Talent acquisition is a human skill. One that requires a deep understanding of who candidates really are beneath cover letters that may or may not have been artificially generated.
Sure, it’s unrealistic to ask your recruiters to gain a deep understanding of every person who applies to every role. But you can empower them with the right tools to give them a head start at identifying top talent. Which in turn leads to better hires, reduced turnover, and an improved bottom line.
Enter: the personality assessment.
Personality assessments get to the truth of what candidates are actually like.
Personality assessments aren't new — they’ve been used to help select talent for decades now. However, up until fairly recently the experience of taking those assessments was long and tedious. Candidates didn’t want to take them and many employers were unhappy with the number of candidates they were losing as a result.
Nowadays things are different. Assessments take three to four minutes. They’re visual. Engaging. Candidates actually complete them at a high rate. And because of AI, they also have a renewed purpose: being “reality filters” in the hiring process.
At Paradox, we’ve made the entire process mobile-first and candidate-friendly: After screening a candidate, a short, visual-based, assessment is automatically texted out. The candidate will look at a series of images and captions, and determine if they relate to each one or not. Because of this simplicity — it only takes about 3–4 minutes to complete — we see a 95%+ completion rate. Responses can then be compared to behavioral benchmarks so that hiring managers can easily see which roles someone looks like a good fit for.
Where does AI fit into all this? Well, it doesn’t. The assessment results are entirely user-driven, and the user can’t use AI to deliver the right answer. Because there is no right answer.
Right now, AI — at least the most accessible models — is very good at making binary connections. Answering yes or no questions. Scheduling interviews when your calendar is open. But when you ask it to infer something a bit more complicated, like say, the complete makeup of a person’s traits, skills, and even flaws, AI tends to struggle a bit.
You might be thinking: “Couldn’t a candidate just ask AI to configure answers that fit the best personality?”
No, not really. That’s the beauty of it: Whereas every company hiring engineers might look for a degree, soft skills are more malleable than you might assume. The “right” soft skills are going to look different across different companies and different roles. A positive attribute for one role or company might be a negative for another. It would be nearly impossible for a candidate, and by extension ChatGPT, to consistently manufacture the right answer.
Because there is no right answer.
Personality assessments aren’t built to be aced like a high school test — they’re built to help your humans gain a better understanding of each candidate. And, by the way, they’re really good at it. AI is not.
The impact of assessments in the hiring process.
All that being said, the value of assessments isn’t limited to filtering real applicants through your job applications.
Where they truly shine is where they’ve always shined, even back when they were long and tedious: helping you hire more people that aren’t just real, but are really good fits. And when that’s happening, the hiring process isn’t the only thing that improves. The entire business does, too.
Personality assessments aren’t built to be aced like a high school test — they’re built to help your humans gain a better understanding of each candidate.
For example, I’ve had the pleasure of working with one of our 5,000+ employee fast casual restaurant clients, Captain D’s. They were looking at assessments as a possible solution to their turnover problem, which at the time was as high as 211%. Together, we researched what made their top performers stand out, and then we created a behavioral benchmark for new candidates based on those top performers.
Assessed candidates that matched the benchmark of high performers were prioritized in the hiring process, meaning manager energy was more often spent speaking to top talent. And those great candidates didn’t just get hired quicker — they didn’t quit. Which turned into real cost savings for Captain D’s.
"Hires that don’t stay with us for two months will each cost us a couple thousand dollars, easily,” said Captain D’s VP of HR Sean MacMillan. “So when we hire the right people, we save that money that we were losing.”-
After two months, Captain D’s reaches ROI on a candidates’ hiring, onboarding, and training costs. Assessments helped lower two-month turnover 21% YoY. Which means because of one extra step, managers are now spending less time talking to candidates who would be more likely to leave, spending more time supporting and onboarding better fits, and shifting investments towards internal development.
“We’re now bringing on people that are suitable for the role, people who are cut out for the role, people who will excel in that role,” said MacMillan. “Our hiring managers aren’t spending time and money training people that wouldn’t be a good fit.”
The truth of the matter is that we live in a reality where it’s easier than ever for a candidate to be “perfect.” AI affords them almost every opportunity to fake it until they make it. Emphasis on that almost.
Assessments are possibly the best way to triumph over this increasingly common AI on AI stalemate in the hiring process. In just a few minutes, they allow for employers to actually understand who candidates are as humans, not AI-burnished resumes. But more than that, they drive real business results for the organizations that use them.
So yes, it probably feels awesome to automate mundane administrative work and move faster than ever. But when candidates are trying to move faster than ever too, you need to switch strategies. Maybe learning to slow down a little — even by just a few minutes — is how you win the war for talent.
