How AI is Transforming Call Center Training
Call Center Management
Call Center Management
The training of call center agents is about to experience a revolutionary advancement. Thanks to machine learning, personalized real-time coaching based on speech-to-text transcriptions and analytics will be made possible.
Typically teams of coaches can listen to only 10 or so randomly selected calls for each agent, however, it’s not uncommon for enterprises to employ 10,000 or more call center agents, whose average call duration is seven minutes. Approximately 11,000 hours of audio (about 300 man-weeks or 75 people) need to be manually reviewed each month. The entire review process typically takes a month to compile the information, create reports and distribute them to the agents’ supervisors. Those supervisors then meet with their agents one-on-one to provide feedback and coaching based on the reports.
Call center agents handle 800 to 1,000 or more calls per month, thus 10 calls per month represent about 1% of an agent’s volume. With such a small sample size, the coaching report is more speculative than objective. Quite often, it’s the reason for a stressful conversation between the agent and the coach. Invariably, the agent cannot remember the call that is referenced in the monthly review. And even if the review cycle is reduced to a week, the same challenge persists.
With the advent of artificial intelligence (AI), call centers will be able to analyze all calls using technology that can transcribe the audio as the caller and agent are speaking. Machine learning will analyze the calls from both the agent and customer side to predict the outcome of those conversations based on a set of goals, hypotheses and other factors. The technology can then measure the predicted outcomes with actual results. Coaching feedback will be provided to the agents in real time while they are speaking with the caller, rather than the existing timeframe of 30 to 45 days after a call is completed.
The ability to do real-time coaching based on real-time monitoring of all calls eliminates the ambiguity from coaching and training, and supervisors can determine coaching opportunities for every agent throughout their work day.
AI not only measures every word spoken; it’s also able to measure the emotional qualities of what’s being spoken, who’s speaking and when that person is speaking, plus it enables the coach to offer needed feedback during a call. For example, the AI engine could determine that an agent needs to convey more empathy to a customer who is considering cancelling a service.
Machine learning that incorporates speech-to-text technology has powerful sales training implications. Take, for instance, a company that hypothesizes sales will increase by 10% if sales agents state during the conversation that the firm is rated A+ with the Better Business Bureau (BBB). Traditionally, agents are pulled from the floor and trained to learn the new phrasing before they’re sent back to the phones to interact with customers using the updated script. The floor manager then waits a week typically for the sales results (expecting a 10% lift). In this particular case, however, sales did not improve.
Most call centers would assume the new pitch was ineffective and go back to the drawing board. But with AI-supported coaching, it can quickly be determined that none of the agents actually mentioned the company is rated A+ with the BBB. When the training coach asks the agents why they’re not mentioning the A+ rating and their response is, “We don't know how to work it into the conversation,” AI can come to the rescue.
AI analyzes all calls and determines when it’s the right time during a conversation to mention the A+ rating. While an agent is talking to a customer, a small notification appears at the optimal moment on his or her screen to mention the A+ rating and offers suggestions on how to announce it. The ambiguity of what’s being spoken by the agent and customer is removed, and results of the new coaching are often available within minutes. Empirical data will show if an agent needs additional coaching, if the modified sales pitch is more effective and if the customer is satisfied.
The coaching is now personalized, with each agent able to maximize his or her effectiveness. What’s more, the AI alerts human coaches to agents who need additional training in order to help improve overall sales.
Harnessing the power of AI, machine learning and speech-to-text technology enables an enterprise to analyze every conversation between every agent and every customer, and determine how to maximize the agent-customer interaction. It provides direct feedback to an agent while the coaching moment is fresh in his or her memory, creating an environment of positive learning and near-instant gratification.
This new personalized training method driven by AI is analogous to having someone walking beside you as you learn to ride a bicycle for the first time. You receive pointers on how to balance, pedal and steer, plus the occasional steadying hand. But without AI, it’s as if there is a person observing through a small window across the street how well you ride your bike and mails a review to you once a week.
The ability to analyze hundreds of millions of calls in a given year means that the AI engines will become smarter over time as they extrapolate and learn from all the data. This will enable the systems to build highly predictive models and scenarios suggesting how agents should interact with customers. The machine learning will then feed the results of these new interactions back into the same data model and revise its approach orders of magnitude faster than humanly possible.
By transcribing all conversations and applying machine learning as the conversations progress, call centers will benefit from personalized and focused coaching for 100 percent of the staff. Training will be continuous, relevant and delivered at the moment it’s both needed and most impactful, with greater proficiency.
Companies that employ AI-assisted training will experience improved customer and employee satisfaction, increased customer and employee retention rates and the ability to adapt to changing market conditions at near-real-time pace.