As companies continue to refine their hiring processes, a new study from assessment firm HiringBranch challenges a common practice: reporting soft skills as isolated scores. The research, discussed in the latest episode of the podcast "You Should Know," suggests that measuring empathy, acknowledgment, active listening, and reassurance separately may weaken the predictive power of hiring assessments for frontline roles.
Assaf Bar-Moshe, Chief of Research and Development Officer at HiringBranch, presented findings that single-skill scoring shows only moderate correlation with human annotators, while a combined proprietary model yields much stronger correlations. This suggests that evaluating candidates on individual traits in isolation may not accurately reflect their ability to handle real-world customer interactions, where multiple skills must work together seamlessly.
"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," Bar-Moshe said during the podcast. He emphasized that HiringBranch takes a linguistic approach, using a "sociopragmatic analysis of the words that the candidate is actually saying" rather than relying on personality-based assessments.
The study focuses on four pillars of customer service: acknowledgment, reassurance through positive language, empathy, and active listening. HiringBranch's assessments use open-ended voice and writing prompts, turning job descriptions into conversation flows that are calibrated per client, region, and role. For example, regional variations across markets like Vancouver, Toronto, and Montreal can lead to different scoring weights for the same position, reflecting local communication norms.
The implications for employers are significant. Many hiring managers rely on assessments that break down soft skills into separate metrics, assuming that a high score in one area indicates overall customer service ability. However, this research suggests that a candidate might excel in empathy but fail to actually solve a customer's problem or reassure them effectively, leading to poor on-the-job performance.
To address this, HiringBranch uses machine learning models built on years of textual data, developed by a team of IO psychologists and linguists. These models predict skills like empathy and acknowledgment, and then are validated against on-the-job performance months after hire. This approach aims to provide a more holistic view of a candidate's capabilities.
The podcast episode, hosted by William Tincup, also touched on practical examples, including a retail confrontation over mispriced broccoli that hinged on diplomacy rather than policy. Such scenarios highlight why isolated skill scores may miss the nuance required in live customer interactions.
Looking ahead, HiringBranch is developing a self-serve capability 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 is set to be published under the AI research tab on the HiringBranch website.
This research arrives as employers across industries, from customer service to retail, are reevaluating how they measure talent against the realities of frontline work. By moving away from isolated skill scores, companies may be able to make more informed hiring decisions, ultimately improving customer satisfaction and reducing turnover. The episode is part of the WRKdefined Podcast Network and is available on the You Should Know Podcast page.
