In our increasingly digital age, businesses are on a constant quest to show customers their “human side”. Whether it’s putting faces to the names behind a company or featuring happy, loyal customers who are “just like you” in advertising campaigns, humanity has become a highly coveted brand quality – particularly for those that have little face-to-face interaction with customers.
However, balancing humanity with efficiency can be a challenge. “I’m sorry, I don’t understand,” doesn’t exactly create a warm and fuzzy feeling, least of all when it comes from a machine recording; but then neither does listening to hold music while you wait for help.
Nonetheless, artificial intelligence has made huge strides in humanizing automated customer interactions powered by instant voice recognition (IVR) systems and virtual assistants. Though somewhat robotic by nature, higher quality data is enabling ever-more fluent customer conversations through speech technology. What’s more, machine-delivered service isn’t always second best, as the popularity of smart speakers has proven. Though we may crave a personalized experience, it seems today’s consumer doesn’t mind so much how they get it, so long as it delivers and fast.
Being waited on hand and foot may be the epitome of VIP service, but when it comes to getting simple things done, speed is of the essence. It’s predicted that by 2020, 85% of customer relationships will be managed without human interaction. If it’s a toss-up between waiting for a polite human to gently guide you through a process, versus the immediate DIY option of talking to a machine, the bot wins hands-down.
The key to success is training the model with the right speech recognition data and advanced natural language processing (NLP) techniques so that customers can feel welcome, valued and understood – regardless of their linguistic register, accent or dialect.
The “right” data means gathering and transcribing voice recordings – whether scripted, prompted or spontaneous – from a relevant and large enough group of people that match the demographic and linguistic profile of a company’s target customer. The transcribed recordings then go through a semantic annotation process, in which sentences are tagged with the speaker’s intended meaning. The data is used to train chat bots or other voice-enabled systems both to understand a customer’s request and respond with appropriate answers or actions.
Understanding meaning through a well-tuned voice recognition model is a huge step towards serving customers at scale, with fast resolutions to common problems and queries. When that model can also detect emotions such as anger, impatience, confusion or doubt, it takes its potential to a whole new level.
Artificial emotional intelligence or “emotion AI” may sound ominous, but in the case of CX its main purpose is about knowing when to pass a customer back to a real human. Through training data techniques such as sentiment analysis, AI applications can learn to detect a customer’s mood from their choice of words, pace, tone and even breathing patterns or pauses that could indicate stress. So before a customer gets irate, they can be swiftly referred to a live agent who can help them out.
The power of AI to add value by understanding people’s emotions goes beyond the front lines of customer service. It also proves invaluable when it comes to analyzing customer opinion and sentiment through reviews, social media and other consumer generated content. The data gleaned from these types of insights can be used to drive a range of actions that enhance customer experience, engagement and endorsement. These actions include acting on feedback to enhance product features, devising cross-sell and upsell strategies, running retention or loyalty programs, or even launching above-the-line advertising campaigns to confront issues and shift brand perception.
Let’s take a step back though. Before customers even start to interact with products and services, brands first need to win their attention, interest and desire. This is no small feat. In the year 2000, the human attention span was said to be 12 seconds. In 2018 it was down to a mere 8. Yes, that’s less than a goldfish. Where digital marketers once strove to turn browsers into buyers, in today’s saturated ecommerce space simply getting someone to browse before they bounce is often a mammoth achievement.
Prospective customers have neither the time nor the will to consider products and services that don’t instantly match their interests. Not only that, if information isn’t presented in a way that matches their learning style (kinesthetic, pictorial or textual), they’re likely to bounce off faster than you can say “pogo stick”.
It’s simple. If you want people to stick around, you need to make them feel at home. That means serving up content, messaging, products and experiences suited to their needs, tastes and desires. The success of entertainment companies such as Netflix and Spotify are testament to this.
Beyond broad segmentation models that place consumers in buckets according to categories they shop in or sites they visit, retailers are now harnessing the power of AI to offer a new dimension of personalization – whether in the digital space, bricks-and-mortar stores, through apps, television or in consumers’ homes. The capabilities of AI to identify, analyze and predict consumer behaviors offer the ability to create entire individualized journeys that increase the chances of getting people through checkout, gaining trust or ensuring they spread the word.
Whatever the end goal, from product sales to app downloads or increased NPS, AI has huge potential to transform the customer experience through enhanced personalization and – however paradoxically – to bring humanity back into the equation.