Leveraging Artificial Intelligence for Customer Service Without Losing the Human Touch

Customers have high expectations when it comes to the level of service they demand from brands. While the American Express Customer Service Barometer found that Americans are willing to spend up to 17% more on businesses with excellent customer service, the top reason most customers switch products or services is because they feel unappreciated by the brand. In fact, 33% of Americans are inclined to switch to a different company after a bad experience.

Unfortunately for companies, the cost of human support is high. Introducing artificial intelligence (AI) into operations is one way companies can control costs while improving their service abilities and maintaining the human touch that makes customers feel appreciated and valued.

What Is AI Customer Service?

While AI and machine learning may at first appear to threaten the customer service industry, they actually have the power to make customer service agents’ jobs less time-consuming and more fulfilling.

Integrated AI can instantaneously retrieve the data an agent needs, while the agent or support team deals directly with the human side of customer service. This eliminates the need for human agents to run multiple systems simultaneously to address customer inquiries. Rather than employ agents to work 24/7 in a call center, AI can be used to field and classify calls and messages so human agents are then able to work more reasonable shifts with increased efficiency and reduced physical and mental stress.

Through intuitive machine learning that constantly works to improve itself, AI allows companies to be present to the very best of their abilities along every step of the customer journey.

How Are AI and Machine Learning Being Used in Customer Service?

There are plenty of reasons why AI and automation should be loved, especially when it comes to customer service capabilities. Here are a few ways the technology is already being used:

Chatbots

Everyone has had the experience of needing a simple question answered by a brand, only to dread having to jump through customer service hoops just to get someone on the phone who may or may not have the answer. Conversational chatbots can make these conversations more seamless. Not only do conversational platforms help cut costs, they also can help your customer service scale and enable your agents to have more meaningful and productive conversations. By using chatbots to aid your live chat operations, your business will be able to engage customers in real time without the need for an around-the-clock staff.

Amazon, for instance, uses chatbots that leverage the data the company collects on all of its customers and their past orders. By allowing chatbots to access information about the customer’s past preferences, you can have the chatbot interact with customers up to the point where an agent is needed. Once the conversation is transferred to an agent, they can pick up where the chatbot left off.

Eventually, you can train your chatbot to not only acquire customer information, but also recommend the actions customers and agents should take next. If a customer simply needs a common question answered about a product they already purchased, the chatbot can direct them to a FAQ rather than contact an agent. This saves the human agent’s time and allows them to make better use of it dealing with more complex customer queries. All chatbot interactions can be automatically tagged in your AI system so they’re easy to track and reference, and can be used to improve future recommendations.

Robotic Process Automation

Robotic process automation (RPA) can be used to handle the necessary, but routine tasks that keep support agents from interacting with customers in meaningful ways. By taking care of low-priority, mundane tasks, RPA helps customer service agents reclaim time in their days that would be better spent handling high-value customers or fully addressing complex questions without feeling rushed.

RPA works across multiple systems to track user actions within an application to complete and perform tasks ranging from automatically replying to emails to routing conversations. The improved efficiency from saved time on menial tasks also saves companies money. Aside from cutting costs, RPA has the power to increase revenue by speeding up the rate at which customers are able to make purchases through your company.

Agent Specialization

In the past, automated phone systems performed data dips, moving customers through a phone tree where they were asked to “press 1 for a current reservation,” “press 2 for reception,” “press 3 to make a new appointment” or something similar. The flaw in this system is that the information collected was never handed off to the agent, and the customer would have to repeat themself once they were connected with a human. AI eliminates this unnecessary process — if a customer is calling about a product that’s discontinued, for example, there might not be a need for a human agent to talk to the customer only to relay that same information. This saves time for both parties by supporting your human customer service agent and saving the customer from exasperation.

Using AI to capture information about the customers and pass along only the absolutely necessary parts of that information allows agents to have more meaningful conversations and become more knowledgeable about the areas of the business that matter.

If a customer still wants to talk to a human even after discovering their product is discontinued, the agent can immediately begin the conversation by offering recommendations for other products the customer may like. AI doesn’t eliminate the need for humans, as many people incorrectly assume when they hear talk of using AI in customer service. Instead, it augments the human team and allows them to be better at their jobs.

Monitor Support Operations

When you use AI to monitor support operations, you can predict when conversations may start to turn from positive to negative. This insight allows managers to intercede accordingly, and no longer requires them to randomly audit customer service calls to regulate quality.

AI can also help monitor which responses result in reopened tickets. If response A, for instance, tends to resolve inquiries quickly, but response B results in the ticket repeatedly being opened, the system can recommend you eliminate response B in order to set your agents up for success. Managers and executives can use the data generated by AI to oversee customer service operations in a more clear, efficient way, improving day to day operations for everyone involved.

What Are the Advantages of Automated Customer Service?

Customer satisfaction is directly linked to the service experience, and so it’s important to make sure the customer journey is as seamless as possible. Integrating AI into your customer service isn’t about replacing humans. Rather, it is about arming your customer service agents with the information they need to have purposeful conversations with your customers, and using data to personalize your customers’ experience with your brand.

Incorporating AI customer service not only improves your relationships with your customers, it builds trust and increases brand loyalty. This means more repeat customers, and more word of mouth referrals for your business.

When you build an incremental strategy to roll out AI in your organization and optimize according to data collected, success is sure to follow. Using AI to build a more complete view of a customer’s relationship with the brand helps companies meet high expectations for exemplary service, and come across as anything but artificial.

Kustomer Offers AI Business Solutions

The Kustomer platform stands out among customer service solutions for the comprehensiveness of available customer data and its business process automation that is driven by branchable, multi-step workflows and custom business logic. Kustomer IQ is a groundbreaking new service that integrates machine learning models and other advanced AI capabilities with the Kustomer platform’s powerful data, workflow and rules engines to enable companies to provide smarter, more personalized, automated customer experiences with increased efficiency.

Kustomer IQ integrates machine learning, natural language processing, predictive analytics, deep learning and multi-dimensional neural network mappings as a part of its AI suite. Natural language processing involves the interactions between computer and human language, and dictates the extent to which computers are able to process and analyze large amounts of natural language data. Natural language processing is used along with text analysis, computational linguistics, and biometrics in sentiment analysis, also known as opinion mining, which helps companies keep a finger on the pulse of their target audience’s interests and values.

Companies that employ the AI suite are then able to use their own data to train Kustomer IQ’s predictive machine learning models, automatically customizing them to address their own business needs. With each new interaction and piece of data, these models learn and self-tune increasing their predictive accuracy and improving the decision making of both the models themselves and the customer service organizations using Kustomer.

Through Kustomer IQ, companies will be able to automate manual, repetitive tasks and essential processes of their customer service experiences. In addition, Kustomer IQ changes the way companies manage knowledge during a service inquiry by surfacing relevant insights and predicting future outcomes to enhance customer self-service, facilitate real time intervention through recommendations, and streamline proactive outreach. By automating everything and providing the right information at the right time, Kustomer IQ frees up agents to focus on more complex and emotional customer interactions, resulting in reduced costs and faster resolution of calls.

Features of Kustomer IQ include automated conversation classification, queues and routing, customer sentiment analysis, automatic language detection, suggested agent shortcuts, customer self-service, conversation deflection and workforce management. If you’re interested in learning more about Kustomer IQ and how it can help elevate your business’s customer service capabilities, download our ebook, explore our website and get in touch today.

Kustomer offers real-time, actionable views of customers, continuous omnichannel conversations, and intelligence that automates repetitive, manual tasks to make personalized, efficient and effortless customer service a reality.

 

Tell Me How You Really Feel: The Best Metric For Finding What Your Customers Need

Brandon McFadden is Kustomer’s Customer Success & Support Manager, you can follow him on Twitter at @brandontonio. Read his post on using CES to help your product and service teams work better together here. The following was adapted from a workshop delivered at Support Driven Expo in Portland, OR.

After recently writing a piece about using CES to help your product teams, I received some questions asking, among other things, what CES even is. So I wanted to go over that here.

Customer Effort Scoring is one of the most effective ways to understand how your audience feels about their experience, and has some distinct advantages over methods like CSAT and NPS. The principle is simple: you’re asking your customers how difficult it was to solve their issue or complete a transaction. Like NPS or CSAT, it only takes one question to get the information you need. Below we can see two examples of CES survey questions:

So what makes a Customer Effort Score such a useful metric? The answer is rooted in human nature, specifically feelings. 96% of customers don’t complain when they’re unhappy, however they’re four times as likely to defect to a competitor if they have a problem. So while finding out if your customers enjoy their experience is critical, it doesn’t always tell the whole story. Here’s the kicker: 70% of buying experiences are based on how the customer feels they are being treated. So even if your service is best-in-class for your industry, if your customers have unknown, higher expectations and your service feels lacking, they’re going to retain that feeling going forward. So the real question for the data-driven team is: How do you quantify feelings?

That’s why CES is so useful—it can tell you how your customers really feel, where other methods focus on intent and how your customers see themselves instead of addressing the feelings that drive their actions. While your clients may give a high CSAT score, what they’re saying is “I really liked talking to your team, they are AMAZING!” (and who doesn’t want to hear that?) but what they might also be thinking (feeling) is, “Why did I even have to call in the first place?” Most people don’t want to speak badly about or hurt the career of an agent, especially when they solved the problem, but they will hold a negative experience against your brand as a whole when their expectation was that the fix should have been easier—or if they never expected to have this problem to start with. To make matters worse, this usually only manifests itself when it is time to recommend your service/product. Lesson? Your agents might be doing great work (of course they are, you hire great people), but that doesn’t always lead to more referrals and repeat customers.

Typically this is where NPS seems like it should provide the other half of the picture you’re missing from CSAT. If customers are satisfied but not willing to recommend you, then something in your experience is lacking, right?. There’s nothing wrong with that assumption, but NPS also has pitfalls of its own, once again sabotaged by feelings. Often, customers will say they would recommend you to their friends, but in practice, they don’t. Interestingly, the problem is found in the NPS question itself: “How likely are you to recommend this product to a friend?”. When we think of our friends, we think of people just like us, same skill aptitude, same patience, same willingness to put up with the “why did I even have to call about this” issues. But in reality, when it comes time to make the actual recommendation, they balk. They think “oh, they aren’t as technical as me” or “they likely don’t have the same patience with that issue like I did”. So while maybe they would recommend your product in general, on a one-to-one basis, they might have lingering doubts about a difficult experience and don’t feel their personal friends would have the patience to deal with your service.

What NPS and CSAT don’t do well is make it easy to identify your customers’ hidden frustrations and reluctance to advocate for you in the real world. Neither help you pinpoint the parts of your product or process that cause the most frustration, not simply have the most quantity. This is why 82% of US companies report that they are “customer-centric”, while only 18% of US customers agree. Clearly, there’s a disconnect between how companies see themselves, and how customers see them. But if their NPS and CSAT scores are high, why should they think otherwise?

Ultimately, this is because customers are thinking: “If you really cared about me, then why are you making it so hard to do something I think should be so easy?” It’s probably a question you’ve even asked yourself when you’ve been on the phone with customer support. Fortunately, with CES, these feelings are able to be captured and quantified.

Let’s look at an example of the Customer Expectation Gap in action. I recently had two experiences where my expectations and the reality were way off, giving me two very different opinions of the organizations I was dealing with after the fact. Those organizations were Amazon and the DMV—about as different as you can get. One is “tech” and optimized to solve your problems, and the other is the DMV.

I’m pretty sure that if I offered you the choice of getting a new license at the DMV or requesting a refund from Amazon—you would choose Amazon every time (and for good reason, their support is fantastic). While I didn’t have to choose in the moment, I did have to get a refund for a Netflix gift-card purchased through Amazon (silly me, didn’t coordinate with my brother). Given their renowned and very streamlined buying experiences, I thought the process would be just as easy. In a way, you could say that they trained me to think this would be just as easy as buying. This, frankly, is the blessing/curse of tech. We spend endless time making things easier, automating, reducing effort—meaning it hurts that much more when this doesn’t happen with Support resolutions. Inversely, around the same time, I needed to replace my license at the New York City DMV—a much-maligned experience and a staple of 90s stand up—albeit for good reasons. I expected this to be an all-day ordeal (ok, maybe half day), because it had been before in multiple states over the past 20 years for me. I had been trained to expect the worst.

However, getting my refund from Amazon was the real bureaucratic nightmare, stretching across four calls and two 15-minute chat sessions, and taking over 2 days to resolve. On the other hand, the DMV was a breeze. I booked ahead online, found an “express office”, checked-in on a screen, followed an express lane to an automated machine, and was done in less than 30 minutes. Now, I’ve been bragging about the NYC DMV to my friends (who think I’m crazy), and certainly haven’t recommended ever getting a gift card from Amazon. The funny thing is that If I had called up Amazon expecting a hassle, I wouldn’t have remarked on it, and if I had known that the DMV had become so cutting-edge (kind of), then maybe I wouldn’t have been wowed. So as you can see, it really is the combination of how I felt about the experience, my expectations, and the relative effort I had to expend that determined whether or not I became an advocate.

To be clear, none of this is to say that you shouldn’t measure NPS and CSAT. You absolutely should, and they are crucial metrics for understanding your business. But if you want to know how your customers really feel about your experience, they leave too many gaps. With CES, you can fill those gaps and get all the context you need to identify where your experience is weak, and how you can improve it. So maybe start by adding a 2nd CES question to your post-issue CSAT survey, you may just be surprised by the results. Remember, it’s not about what your customers say—it’s how they feel that creates impact at the moment of their referral, making repeat purchases, and when they decide to churn. If you would like to learn more about how you can act on this information, feel free to check out the companion piece: How CES Can Help Your CX and Product Teams Work Better Together.

To learn more about how Kustomer can help you better understand your customers, request a demo below!

Deliver effortless, personalized customer service.

Request Live DemoStart Interactive Demo