As employers increasingly rely on pre-hire assessments to screen frontline talent, new research from HiringBranch suggests that a common practice—scoring soft skills separately—may undermine the accuracy of hiring predictions. The findings, discussed in the latest episode of the podcast You Should Know, challenge the conventional approach to evaluating empathy, acknowledgment, active listening, and reassurance in candidates for customer-facing positions.
The episode, titled Assessing Skills One at a Time Is Costing You Better Hires, aired August 26, 2026, and featured Assaf Bar-Moshe, Chief of Research and Development Officer at HiringBranch. Bar-Moshe, a trained linguist, explained that their open-ended voice-and-writing assessments, which simulate live customer interactions, yield stronger correlations with human annotators when using a combined proprietary model rather than individual skill scores. “If a candidate can express empathy, but is unable to solve the issue correctly or to comprehend the issue correctly or to reassure the customer, then this empathy is nice, but it's actually useless,” he said.
The study zeroes in on four pillars of customer service: acknowledgment, reassurance through positive language, empathy, and active listening. Rather than isolating these traits, HiringBranch’s model integrates them into a holistic evaluation that better mirrors the demands of real-world interactions. Bar-Moshe described the approach as a “sociopragmatic analysis of the words that the candidate is actually saying,” distinguishing it from personality-based assessments.
The implications for hiring are significant. For roles like customer service representatives, sales agents, and retail associates, the ability to navigate complex emotional situations is paramount. Host William Tincup illustrated this with a retail confrontation over mispriced broccoli, where diplomacy mattered more than policy. Such scenarios underscore the need for assessments that capture the interplay of skills rather than isolated traits.
HiringBranch’s methodology involves converting job descriptions into conversation flows and scenario-based assessments calibrated per client, region, and role. The company’s team of IO psychologists and linguists uses years of textual data to train machine learning models that predict these skills, validating predictions against on-the-job performance months after hire. Notably, calibration varies by geography—for example, scoring weights differ across Vancouver, Toronto, and Montreal for the same role.
Looking ahead, Bar-Moshe previewed a self-serve capability in development that would allow hiring managers to build assessments from a library of conversation flows and skills, reducing reliance on weak or generic job descriptions. The full study will be published under the AI research tab on the HiringBranch website.
This research arrives as companies rethink how they evaluate frontline talent amid evolving customer expectations. By highlighting the limitations of single-skill scoring, HiringBranch adds to a growing body of evidence that integrated, context-aware assessments may better predict job success. For HR leaders, the takeaway is clear: measuring soft skills in isolation could mean overlooking candidates who excel in live interactions, while overvaluing those who score well on narrow metrics. The shift toward combined models, as suggested by this study, could refine hiring practices across industries.
