Do The Best Coaches Combine Man And Machine?

Coaching and mentoring have seldom been more popular, with a growing consensus that it can be vital to our career development.  Traditionally, the role of the mentor has been filled by an experienced colleague who can provide us with the wisdom they’ve gained over their careers.  As new research from Temple University reminds us, however, it needn’t always be that way.

The paper highlights the growing use of AI to coach workers, especially in the sales domain, where AI coaches analyze sales agents’ conversations with customers and provide feedback to them on their performance.

While it may seem somewhat unusual, Zoom has Chorus, its AI coach, as a standard feature, and products like Cogito have also been working extensively with call center staff in recent years.  The allure of such products is that they’re capable of crunching the huge quantities of data generated during the course of this kind of work to provide informed feedback to employees and employers alike.

Combined feedback

Despite this processing heft, however, such tools are generally only good at providing “hard”, data-driven feedback, so the researchers advocate pairing them with human coaches who can provide feedback on soft skills.

The researchers argue that this lack of soft skills may even ensure that the employees don’t fully absorb the feedback they get from the AI coach either.  This can be exacerbated by the fact that many AI tools focus on information generation rather than learning.

They conducted a number of field experiments in a couple of fintech companies.  For instance, in one, a few hundred agents were assigned to either a human coach or an AI coach for on-the-job sales training.  The results suggest that the AI coached produced an inverted-U shape outcome, with middle-ranked agents improving but bottom- and top-ranked agents showing limited gains.

The data suggests that this pattern is largely driven by the underlying mechanism of the tools themselves, as the bottom-ranked agents suffered from information overload.  By contrast, the top-ranked agents tended to have an aversion to the technology itself, which obstructed their learning.

This, of course, matters, as those lower-ranked agents have the most capacity for learning.  The researchers tweaked the mechanism and conducted a follow-up study that restricted the amount of information given, and this seemed to improve outcomes.

Man and machine together

The best results were achieved, however, when man and machine worked in unison.  When human coaches worked alongside the AI-based tools, there was more significant improvement across the spectrum.

The researchers believe this is because such an approach combines the data analytics power of the tech and the soft communication skills of the human coach to provide a compelling combination.

“Managerially speaking, our research empowers companies to tackle the challenges they may encounter when investing in AI coaches to train distinct types of agents. We show that instead of simply applying an AI coach to the entire workforce, managers ought to prudently design it for targeted agents,” the researchers conclude. “Companies should be aware that AI and human coaches are not dichotomous choices. Instead, an assemblage between AI and human coaches engenders higher workforce productivity, thus allowing companies to reap substantially more value from their AI investments.”

Facebooktwitterredditpinterestlinkedinmail