Artificial intelligence (AI) innovations in the customer service industry are accelerating. AI applications are common in the customer service space, such as chatbots that have replaced some of the duties of customer service representatives. The impacts of its applications are far-reaching and have also made waves in the training of customer service professionals. 

Some of the most promising and sophisticated applications of AI in customer service complement the skillsets of subject matter experts (SMEs) to maximize their impact. For example, AI-based coaching software, like Zenarate, Cogito, and Nuance, have carved a new niche into the customer service training process by democratizing accessibility to training coaches instead of overburdening team SMEs. Products in G2’s Proactive Customer Retention category use AI to predict trends and report complaint patterns for awareness among customer success teams. However, a new AI-driven training program for customer service professionals is sorely needed in the industry.

A virtual role-play training experience for customer service professionals

Onboarding new customer service professionals can be a time-consuming and expensive process for any business with a large customer base. Newly hired employees are regularly expected to start performing on the job as soon as possible, which can negatively impact the quality of their customer service. Turnover in these types of roles can reach up to 45% every year, with some call centers reaching triple digits. These trends further illustrate the need for a robust and effective training curriculum for new professionals to get them up to speed as soon as possible. 

Proper training also directly relates to employee development, which further impacts job satisfaction and retention rates. One way to deliver this training experience is through job coaching using products usually found in the Contact Center Quality Assurance category. However, coaching is a time-consuming and cost-intensive process, as it both takes a coach to train and can prevent high-performing employees from performing their standard work duties. On the job, experiential training is also costly in regards to customer trust because learning from mistakes can come at the expense of the customer’s experience.

So what can an AI-coaching solution do to help your customer service team? The AI-coaching tool approaches automation differently from most AI products by focusing on developing employees instead of replacing them. To accomplish this, it steps into the role of an onboarding training coach and partner to build skill sets and expertise. 

As a training partner, the software can participate in customer engagement and imitate the role of a customer. These engagements are designed to expose the trainee to situations and inquiries they need to be familiar with to excel in their role. They can range from basic company policy queries to preparing representatives for emotionally charged situations requiring de-escalation. Using the data gathered from each of these engagements, the AI-coaching solution comes forth and analyzes their performance to provide feedback. As a coach, the product tracks performance and provides feedback during training sessions.

AI limitations: Considering user experience and emotional intelligence when delivering feedback

The technology is still new, and users will likely respond to an AI coach with hesitation versus a human counterpart. The effect of the uncanny valley is a well-established phenomenon where the closer the AI avatar gets to human likeness, the more difficult it is for the user to accept. For example, robots that emulate human emotions are often viewed as unnerving. As such, research is critical for ironing out these kinks in the user experience.

Source: ResearchGate

In a study on how AI-coaching software could improve the performance of sales teams, experimenters found two factors that predict the product’s effectiveness when delivering feedback: information overload and emotional intelligence. 

The researchers found that when they divided the experimental group into the top, middle, and bottom (newest employees) performers, the group that improved the most from AI coaching was the middle performers. This finding was problematic because businesses that adopt this software often expect their lowest performers to benefit the most. Intrigued by the finding, researchers believed that bottom performer users weren’t responding well to the feedback because they were experiencing information overload. To compensate, they lowered the quantity of feedback from a list to a single suggestion predicted to be the most impactful in improving their performance. When these changes were made, the bottom performers started improving their performance the most.

Finally, the researchers compared how the source of the feedback impacts performance. They compared AI, human, and AI plus human combined coaching conditions and found that having the software identify the areas for improvement in employee performance while a human messenger delivered the feedback (AI and human coaching combined) showed the most improvement. 

Artificial intelligence will continue to struggle to interpret social and non-verbal feedback when interacting with humans. For example, the telltale blank stare people wear when they receive a torrent of performance feedback usually prompts a human manager to slow down their delivery style. Humans know that delivering feedback in an emotionally intelligent manner encourages the trainee instead of discouraging them. For now, AI can only be as considerate and socially adept as they are programmed.

Looking forward

AI-coaching software is promising and very powerful, but studies into human-computer interaction show that it is still best applied with a human touch. The human-AI interaction compatibility gap remains wide, and building the bridge across that uncanny valley will take time and research. Until then, the innovations and paradigms being developed build those critically important foundations of said bridge.

Edited by Shanti S Nair