Large companies are more eager than ever before to better understand their customers. We want to anticipate what our customers want and what they need so that we can better serve them and earn their loyalty and trust.
The problem is that traditional customer interactions simply don’t capture this kind of data in a way that is usable and scalable for a large business. You can’t just poll your call-center agents. Inherent biases render the questions you ask as suspect as the answers you might receive. You can survey customers, but you can’t conduct enough surveys to keep up with the constant evolution of your business, consumer expectations, market conditions, and other variables largely beyond your control.
But the emergence of smart machines like bots, IVA’s, and even simple web-chat tools not only give us the means to capture this data but also do it in a way that lends greater context and meaning to analyses of the customer experience. Here’s how you can use AI-powered tools to better understand your customers.
Go beyond the “what”
The analytics on your site might tell you what support articles are most accessed by your customers, but they don’t tell you how frustrated the customer is or how they got to your site.
We fixate on the “what” in customer experience. We want to know “what” the customer wants, but a great customer experience accounts for the how: how our customers want to access support, how they feel about particular problems or issues they’re experiencing, and how we can we deliver a customer experience that better meets their needs.
It’s misleading to assume that bots and chat systems can provide value solely by automating parts of the customer experience. The most sophisticated systems add immeasurable value when deployed not only as tools for automation but also as listening posts that you can use to capture and analyze conversations with your customers.
In contact centers we traditionally measure time to resolution as the driving KPI since human agents are much more expensive than software. By incentivizing agents to accelerate the process, we introduce bias into the interaction. We often ask customers if their issues were addressed “promptly.” The question itself carries a bias that fast service is more important than anything else. What if the customer really wanted friendly service, or expert service?
Furthermore, agents are generally more likely to record a more positive outcome from the customer interaction since their performance is judged on that basis. Even if we ask customers to complete a survey about the agent after the call, their opinions of the agent may be colored by the product issue they were experiencing in the first place. You get bias from both sides.
Machines, on the other hand, can work towards a balance of competing KPI’s and have no ability (or reason) to skew the reporting. Even in the event of a call with an abusive customer, the machine is unaffected.
The result is the most candid, measurable view into the customer experience that you’ve ever had, which allows you to more accurately assess whether you’re really meeting your KPIs.
Don’t just capture data, capture human data
It’s impossible to discern a customer’s emotional state when they access a support document or search your site. Conversational AI captures structured and unstructured data that can reveal customer sentiment, tone, and intent. When you overlay that “human data” with product lines, time signatures, and other internal data, you can unlock remarkable new insights about your business processes and products.
IBM’s Tone Analyzer is one great example of a technology that does just that. It can be used to recognize when a customer is upset and deliver more empathetic responses when answering their questions, for example.
Consider the recent issues with the Samsung Galaxy Note: users
were outraged and frightened by the sudden explosion of their phones. But the Note is just one product in a massive consumer portfolio of mobile devices, monitors and TVs, appliances and wearables. These products all play some role in the lives of real users, but most companies are blind to the customer’s emotional relationship to the product. We rarely design feedback loops that identify what customers love most and what frustrates them. All too often, we simply make inferences based on the sales figures of a given product to determine how well-liked it is by consumers.
Now consider the tone and sentiment data that bots and other smart systems can capture from interactions with these customers. Without any major overhaul of the customer experience, even the largest companies can now gain a granular view of customer sentiment, likes and dislikes overlaid across sprawling product portfolios.
Embracing Machine Intelligence
Despite the irony, it’s no surprise that machines are the best way to create a more human understanding of the customer experience. Business operates at great speeds and scales these days, and every customer is unique. If you truly want to understand your customers, AI-powered machines are the very best listeners you can employ.
Sure, they can also resolve problems for customers, take actions on their behalf, and even entertain them a bit in the process. But don’t overlook what happens after that interaction. The resulting data is a force-multiplier on the technology's business value.
Traditional customer support is, by design, reactive, which is why it’s dominated by a focus on time to resolution. The new paradigm of customer experience is proactive, and it requires not only anticipating what your customers need but also knowing how to deliver an excellent experience on-demand and at scale. That’s a tall order. Thankfully, machines are here to help.