Interaction Analytics Helps Improve Coaching/Training
Technology vendors today are bringing customer interaction analytics systems and agent coaching/training systems together to identify problems during interactions and then guide agents in real time to overcome them.
Though the systems are still largely separate, they are working together more and more as well as evolving on their own to include more data touchpoints, analysis, artificial intelligence, and machine learning as companies look to provide better customer support and customer experience (CX).
“There is undoubtedly growing momentum around bringing real-time analytics and agent coaching together to drive meaningful customer experiences,” says Rachel Lane, contact center solution principal at Medallia.
She notes that agent attrition is a critical factor behind the trend. “With many contact centers struggling to hire to cover frontline attrition, there is huge pressure to get new hires onboarded and ramped up as quickly as possible. Real-time agent coaching triggered by conversational analytics running alongside an engagement can be of great benefit to help newbie agents.”
HOW IT STARTED
The trend started in the 1990s with the first wave of artificial intelligence, according to Shahar Chen, founder and CEO of Aquant, a provider of prescriptive analytics and intelligent triage technologies for service operations. “It started with expert systems. Companies would take their best people, sit them in a room for six months, then try to download whatever these people had in their minds.”
While designed to share essential knowledge and keep it from leaving when a person left the company, this practice was costly, Chen explains. “You lost your best people. They were no longer in the field. You had a lot of expensive resources sitting in a room for six months. Rather than solving customers’ problems, they were sitting in a conference room. The result was that once you got down to what people had in mind, some of that information was no longer relevant because the models and the machines had evolved. Software updates had come out, too.”
A solution for many companies was to hire professionals who were more seasoned so that they could hit the ground running. But they came with a price, too.
Chen points to three factors that create problems in this regard:
- The pool of experienced professionals continues to shrink.
- The experienced professionals could command better pay or better working conditions, sometimes by moving to a new company, so companies could no longer count on retaining people for years or decades.
- Technology continues to evolve: With artificial intelligence and machine learning, today’s technology can learn from its mistakes. Today’s systems continue to get smarter and smarter.
That’s why adding analytics into the coaching and training mix is so important.
“By augmenting your coaching and performance management with analytics, you can create better agents and even better CX outcomes,” says John Thompson, head of sales at Ibex, a digital customer experience outsourcing firm. “Investment in coaching and upskilling creates a culture of dedicated, highly skilled employees who deliver better, faster results. Invested agents stick around longer, ensuring valuable CX knowledge doesn’t exit your program, leaving gaps in your CX delivery. Better agent development reduces turnover and keeps program expertise embedded in your system.”
But there’s a right way and a wrong way to do this, according to Chen.
“The right way to augment people’s experience is by looking back at the data. Companies understand today that the experience that people have is documented somewhere in the organization,” he says.
“The relationship of coaching/training and analytics systems is based on the idea that agent training/coaching systems learn directly from the customer analytics systems,” says Trey Norman, chief operating officer of Mindbreeze. “They have the ability to combine practically into one system.”
Connecting these two systems enables high-level data extraction that helps the entire workforce complete tasks across multiple functional areas, not just to boost performance in the contact center, Norman adds.
Analysis of the data from human and bot interactions with customers then can inform agent training systems to help boost human performance, Chen agrees. “You can democratize experience across the entire organization. When that happens and then you hire somebody [new], the time to competency becomes very, very short. You can hire someone with almost no experience and they can start being productive super quickly. That solves the initial problem—the inability to hire experienced people.”
HOW IT WORKS
Agent training systems look at the language used in emails, online support, and phone conversations to understand the weaknesses and strengths of specific strategies, Norman adds. “These insights can then be used to directly train and optimize the capabilities of all employees in a customer-facing role. Natural language processing (NLP) drives this method. It ensures that systems are truly understanding human language, [including] verbiage, tone and even sentence structure.”
The systems continue to evolve, so more insights are available every year, moving agent training systems closer and closer to having a 360-degree view of the customer, Norman says. The newer insights are in the realm of AI.
Mindbreeze has focused most of its development efforts in the past year on helping its customers gain more insights from the information scattered in the different systems across an organization, Norman says. “We’re growing the out-of-the box integrations that we have.”
With the data and analysis readily available organization-wide, it is much easier for different trainers to work with different frontline agents, even if they haven’t worked with them before, Norman adds. “The goal is to always have deeper and deeper understanding of interactions so you know what was discussed to inform the coaches so they can better guide the agents moving forward.”
With AI and no-code tools, companies can better classify each customer as well as each customer interaction, Norman adds. “The deeper insights that you can get from these analytics, the more guidance you can provide to agents. That’s where the growth is really going to come from.”
BEST PRACTICES
Companies need to understand that there is no single system that can act as a source of all truth because different systems contain different information, Chen says. “You need to look at all of the data together in a holistic view to understand the [customer’s] full story. When you take the data from all these systems together, then you can understand the right way to solve the customer’s issue.”
To be successful in using interaction analytics systems to inform agent training systems, there needs to be internal support from the team that will actually be doing the work, Chen says. Lack of support is usually due to poor communication. Often there’s a misperception that AI is coming to replace people rather than help them perform their jobs better.
But it’s not just the customer service worker who stands to gain from the merging of analytics and training.
“I’ve led teams that deliver analytical insights that put humans at the center of their design, helping them ‘automate the ordinary so they can focus on the extraordinary,’” says Michael Gill, vice president and chief data analytics officer at EIG, parent company of Employers Insurance. “To place customers and employees at the center of design, strategy and analytics leaders must first holistically understand employee workflows and processes, as well as the customer journey.”
There are moments during the customer journey when they either make decisions or require guidance, such as the point of purchase, Gill explains. “These moments are fortified when a company can provide a more guided experience, informed by employee expertise that is augmented by analytical insights.”
Companies that provide their account executives with specific, actionable, data-driven insights about a prospect lead to more tailored offerings and, resultantly, a boost to sales conversions., Gill adds. “One way to master this and to incorporate a human-centered solution design is by embedding insights generated by a customer conversion predictive model within a CRM.”
Employees—specifically those in customer success roles or in charge of maintaining customer relationships—need access to those insights inside the applications they already use, Gill adds. “To deliver on these experiences, it’s incumbent on chief data analytics officers and the talented practitioners within their organizations to design analytical insight solutions so they can be easily adopted within employees’ workflows, instead of expecting employees to figure out how to access and leverage them in clumsy and cumbersome ways.”
As much value as there is in using customer analytics systems to help inform agent training systems, some companies fail to benefit because they look at the wrong metrics, Norman says. For example, they might be focusing on the length of customer interactions rather than the success of those interactions, Norman says.
Another challenge is the ability, or lack thereof, to analyze exactly what is being said and by whom, according to Medallia’s Lane. “There are typically two ways of triggering real-time coaching through conversational analytics: by chat, which offers high levels of accuracy, and by speech analytics. Speech analytics can deliver high levels of accuracy, but it has the added complexity of call quality, and quality can be affected by the environment that the calls are made in. Contact centers often provide a lot of background noise, giving acoustic challenges, but the same is also true for the customer who can be calling in from a noisy environment.”
The next challenge is the repository in which coaching notes are kept, Lane adds. “Organizations must ensure that they are kept up to date with the very best advice and that all links to associated forms and documents are equally available.”
This stage is the “moment of truth,” Lane says. “The minute that information served to the agent is not absolutely correct, especially the first time, it is likely the agent will lose faith in the materials and is likely to dismiss them in the future.”
These triggers then become a call irritant and their value will be lost, Lane adds. Real-time coaching is a great addition to the agent toolkit to ramp up success, but it needs to be well maintained and monitored to ensure ongoing success and deliver on the promise that it will serve the agent well.
“Everyone knows that the data is important, but the timing is as well,” says Emily Gray, chief customer officer at Playvox, a provider of contact center quality assurance software. “The closer that you can bring the coaching to the actual event that is happening, the better.”
So eliminating gaps in delivering that information as quickly as possible is essential for companies wanting to gain the greatest benefits from aligning customer interaction analytics technology with agent training systems, Gray says. “Having real-time data is the first challenge for some organizations. Then it’s having real-time coaching conversations to promote the intended actions.”
The real-time coaching aspect has become a bigger hurdle for many companies, particularly as agents went remote with the COVID-19 pandemic, Gray says. Though many workers have returned to the office, there is still a large percentage of remote agents, and that is unlikely to change for some time.
But there is a way to get around that challenge.
“By pulling in existing dashboards, you can bring the numbers closer to where the conversations are happening: coaching sessions or team huddles,” Ibex’s Thompson advises. “The data can also be used to automate the process of setting goals and setting trigger points that would automatically generate a coaching request.”
Thompson adds that no two agents, though they take many similar calls, are the same. So by infusing technology into the coaching systems, agents can become much more comfortable with a wide array of situations they could face every day.
“Another challenge is that customer expectations have grown,” Gray adds. “And customer-agent interactions have become more complex. So that in turn has made coaching conversations more complex. Some of the brands that we work with are still struggling to see how coaching fits in with the culture of their organization.”
Gray says that some companies have evaluation issues that can affect training as well. Evaluations from quality leaders and from customer scorecards can be quite different, providing different indications of what kind of training is warranted. If such differences occur, Gray recommends re-evaluating the questions that quality leaders are asking.
But of all the emerging technologies and strategy revisions, no one single element is expected to have more of an impact on both coaching and intelligence than AI. In the next year, AI will push coaching systems to the next level, Chen says. “Those technologies are going to enable us to get to the next level of sophistication and productivity.”
Norman agrees, pointing to recent advances—like OpenAI’s ChatGPT—becoming incorporated in customer interaction analysis and agent training systems. The analytical systems will become more adept at providing summaries, he says.
And with AI, the possibility of human error is removed from the equation, which could positively alter both the analysis and the recommendations needed for agents and customers to get along optimally.
Phillip Britt is a freelance writer based in the Chicago area. He can be reached at spenterprises1@comcast.net.