Recently we met with an executive who shared that his organization intended to go from zero use of chat bots to one hundred percent in an effort to impact customer experience. No escalation path, no testing phase – simply launching new technology in an effort to improve service. While it’s no secret that Humach is a proponent of the power of humans plus machines (it’s in the name), there is a strategy to use and a methodology to apply when considering how and when to use technology.
“There is a framework that must be developed. The data must be structured and indexed, the front-end must be programmed, quality assurance is a must and the whole program must be tested,” says Tim Houlne, CEO of Humach. “You don’t just flip a switch and go from zero to one hundred.”
Technology is an integral part of our daily lives, so it’s easy to look toward a technological advance as the potential solution to a challenge. However, as this HBR article explains, technology must be used appropriately in order to improve customer engagement. In spite of the focus on technical innovation, there is still an important customer aspect that requires understanding the unpredictable nature of human behavior. If you work on a customer experience strategy that balances the two, then good things will happen. If not, then you will have problems.
For example, the article discusses how a bank instituted better ATM and online banking technology, which increased the number of transactions patrons were making, thereby increasing their visits to the brick and mortar bank, and ultimately costing the bank more money. In another example, a major insurance company opted to automate their condolence greeting when customers called in regarding a life insurance policy. It’s no surprise this has not been well received. Rather than taking technology to the extreme, and potentially causing problems for your organization and your customer, Houlne recommends keeping it simple.
“The ideal scenarios to automate are simple decision trees where you know your total universe of variables. If there are unlimited variables or answers that require a decision based on the input, leave those with the qualified agents to handle.”
Great Customer Experience Requires Human Intervention
Customers often prefer to self-serve, making it difficult to know when to automate and when to involve a qualified agent. The innovation of self-service makes automation appealing; however, it can also create pitfalls.
Eliminate the guesswork of customer experience design. Read our e-book!
“I believe one challenge is the confusion that lies around the differences between chat bots, machine learning and augmented intelligence,” says Houlne. “Chat bots are really more of a front-end to the automation that supports machine learning and natural language processing, which leads to augmented or automated intelligence. Even with these advances, there is not a machine that can replicate or replace the human brain.”
The automation of customer experience requires good data in the form of well-organized knowledge bases and frequently asked questions. That data must be indexed and structured to be accessed and delivered appropriately to the technological interface. Only then can you even consider implementing automation to serve up answers and provide customer engagement.
“One of the big challenges is the disparate data within an organization. Some of the pertinent information might by in a CRM system, some of it might by in marketing software, and some in the accounting programs,” says Houlne. “You first have to tackle how to compile all those data sets. Only then can you hope to have your automated systems extract the right information to provide answers.”
“Our first recommendation is going to be to ensure that your foundation for data is strong before you attempt to embark upon any automation. This will require a large amount of testing and organizing and requires the intervention and support of educated agents who can sort out good answers from bad.”
Humans Plus Machines Make Great CX
We know that savvy human agents are needed for complex issues as well as empathetic answers. It turns out that these agents are also required to ensure that your automation delivers good customer experiences. A simple internet search will provide you with plenty of fodder for why you cannot leave automation to its own devices. It’s more than just providing the initial data. Just like training great agents, automation requires its own level of training and consistent interface to actually provide the great service an organization desires.
“To implement any automation requires a crawl, walk, run strategy, rather than a simple flip of a switch,” says Houlne. “You begin with the more simplistic transactions, anything that can be broken down in limited answers and variables. It won’t be one hundred percent, but maybe seventy percent. And you must ensure your team has everything they need to edit and review the automated responses.”
To empower the team to successfully implement automation goes beyond aligning the data sets. It also requires a review of the policies, processes, procedures and a solid quality assurance / quality management system.
“Customer experience scores are often low because of outdated policies and training related issues. Automation won’t fix those and, in fact, will only shine a light on them, making the customers and agents more frustrated,” says Houlne. “Quality assurance and quality management is a critical component to uncovering internal issues that must be resolved.”
The staged approach to automation will uncover some of these internal challenges, providing opportunities to update old policies and provide more training to close gaps.
“If an organization has a proper, frequently updated customer journey map, all of this information will be addressed and the crawl to walk phase will be a lot shorter,” says Houlne. “Often we see organizations with a journey map that is static and hasn’t been updated since it’s completion several years ago. The customer experience process is fluid and must be continuously monitored and updated. Companies must be working to accommodate all of the new stages on that map, including testing and analyzing automation.”
The Humach team has experience supporting clients through the crawl, walk, run phases of automation implementation and uncovering challenges in training, processes and procedures that inhibit great customer engagement. Through Humach Labs, organizations can incubate ideas to get a solid feel for the impact to customers, agent, and the organization overall before they roll it out. Additionally, the Humach agents are adept at supporting clients with the most advanced issues while providing feedback on opportunities to improve.
Contact us today to explore the best way to use automation inside your organization to actually improve customer experience.