Three Steps to Acing AI-Human Interactions


In 2018, the Harvard Business Review wrote about how artificial intelligence (AI) and humans are “joining forces” across many industries. AI has gotten good at doing many jobs that we previously thought only humans could handle: “diagnosing disease, translating languages and providing customer service—and it’s improving fast,” they asserted.


Over these last two years, we have seen the number of these AI-human interactions rise significantly, in large part because of the improvements in conversational AI performance and an increasing number of organizations buying into the promise of all AI can deliver. In a contact center setting, your customers might engage with conversational AI platforms in a variety of ways:

  • Interactive voice response (IVR) to provide telephony-based self-service and customer support.
  • Chat or interactive text response (ITR) to provide text-based customer support, such as live chat on the web or by SMS messaging.
  • Customer service messaging for two-way assistance asynchronously or in real-time across a range of platforms, such as social messaging.
  • Virtual and personal assistants, which rely on natural language understanding (NLU) to handle more complex inquiries by engaging customers in intuitive, intelligent conversation.

Today, AI is part of many brand interactions, but that doesn’t mean those interactions are always successful or add value to users. It has never been more critical to deliver a customer service experience that engages your customers with the right message, at the right time, within the right channel. Recent research reveals that there is a “direct correlation between AI adoption and superior business outcomes.” Among the companies that participated in that survey, researchers found a nearly 25% improvement in the customer experience.

But there are still barriers to consumer acceptance of AI-powered interfaces. Another study revealed that most people still prefer to interact with human agents. Moving the needle in this area will require companies to redefine customer engagement and deliver uninterrupted and seamless customer experiences across all channels (phone, live chat, messaging and so on) and—more importantly—ace those interactions.

Transforming the Journey with Digital Customer Care

To do that, successful companies must “embrace collaborative intelligence” and consider a three-step approach to transforming the customer journey into an experience that delights consumers.

1. Identify Pain Points

Before applying technology to solve customer service challenges, begin by understanding where, why and how often AI-human interactions go awry. Every organization’s experience will be different, but will require in-depth exploration of your agent-assisted interactions and automated engagement alike.

For example, “chatbot fails” can range from not directing or misdirecting to an agent to misunderstanding the inquiry and providing an inaccurate or irrelevant response. Likewise, we’ve all dealt with an IVR system that seemingly runs us around in circles, never connecting us with the information or representative we need. (How many times have you just said “representative” over and over?)

These breakdowns can happen because the technology was deployed before it was ready, with programming errors that snuck into production. But as McKinsey has written, the technology may not be to blame at all. That is, the platform may have been developed primarily to cut costs rather than to improve the customer experience—an issue that creates inherent problems right from the start.

Getting to the heart of the problem will require plenty of analysis and possibly some soul-searching along the way. Ultimately, however, it’s important to remember this truism: Customers are engaging with your agents (whether live or virtual) for one reason—they need an answer to their question or a solution to their problem, and they want it quickly, which brings us to the next step in our process of acing AI-human interactions.

2. Brainstorm Solutions

Once you’ve identified the problem(s) and their source(s), take the next step by brainstorming how to solve those issues while meeting customers’ needs in an efficient, frictionless fashion.

For years, much has been written about creating a frictionless customer journey, one in which your customers can effortlessly meet their needs and answer their questions. But if customers call into your contact center and wait on hold for what they think is an unreasonable amount of time or have to re-answer authorization questions they’ve already entered, that’s far from frictionless. In today’s instantaneous, on-demand world, customers’ expectations are higher than ever.

For instance, we know that customers move between devices and channels as they interact with your brand, and they expect every touchpoint to feel as fluid. Perhaps they start on your brand’s website and interact with your live chat before dialing into the contact center and engaging with the IVR, all before connecting with a live agent. If your customers must repeat information or start the inquiry process over at any point, their experience will be frustrating and may even cause them to abandon your brand, especially if that interaction was started by the brand itself, either via a notification or outreach communication.

The solution for one organization will be vastly different than it is for another. That’s why the key is brainstorming a wide range of possible solutions for creating efficient and effective customer interactions that solve the pain points identified in the first step. Your solution might be as modest as augmenting your legacy IVR with speech recognition or as comprehensive as implementing a digital, omnichannel customer care package—or somewhere in between. In any case, your brainstorming list should include solutions that put your customer first, focusing on delivering account support, service and retention with first-contact resolution and lifetime customer value.

Once that brainstorming list has been developed, you can begin to refine it by prioritizing which options are ideal for your customer base, organizational goals, technical capabilities and industry. Once you have discovered your best path forward, you’re ready for the next step: designing your AI-powered solution.

3. Design and Develop a Conversational AI Strategy

The key to mastering AI, of course, is design. And that begins with a conversational AI strategy that becomes the catalyst for success in customer experience. Creating a customer experience infused with conversational AI every step of the way is a journey and one that shouldn’t be taken lightly.

Begin with your vision, goals and objectives, as well as a thorough understanding of the brand experience your customers expect and that you want to deliver. Then, consider the types of solutions you brainstormed in the previous step, and identify the key user experience steps needed to enable them. Should the solution assist customers with self-service activities? Answer frequently asked questions? Create customer acquisition opportunities? Integrate all communication channels? The possibilities are nearly endless, and only you can know what the right answers are.

Once you have that solid foundation, you can begin to weave in the AI-powered technologies that can support your digital transformation. For example, your contact center likely already relies on an IVR platform to help customers navigate a telephone menu system and route their calls to the right agents. Augmenting and modernizing this platform with conversational AI, however, can significantly improve the customer experience. By allowing the IVR to listen, understand and respond to callers, the system will anticipate customer needs, proactively providing information and rapidly resolving issues without engaging a live agent. In short, a conversational IVR helps alleviate frustrated callers, decrease IVR abandonment and support customer retention efforts.

Having a conversational AI strategy at the ready will mean more than that, though: It will mean your augmented IVR platform isn’t developed in a silo. Instead, you’ll have a roadmap for the future of your business, preparing you for an omnichannel customer service platform—one that provides a seamless, frictionless customer experience as customers interact with your brand before, during and after that IVR experience.

Delivering Extraordinary Digital Engagement

Now more than ever, consumers expect immediate assistance with hyper-personalized advice regardless of how they engage with your brand. Following these three steps to acing your AI-powered customer interactions can get you there, but at the same time, you can’t simply “turn on” the AI and let it be.

Once you have your solution(s) in place, take advantage of the platform’s analytics, which will help you understand what your customer interactions really look like and how they evolve over time, giving you greater visibility into engagement, changing patterns, nascent opportunities, and the root cause for both successes and failures. In these ways, you can easily evaluate your customer experience to identify both best practices and areas for improvement, creating more successful AI-human interactions than ever.

Eduardo Olvera is the Director of Conversational Experience Design in the Technology Advancement Group at Nuance Communications. His work on many international English, Spanish and French voice user interfaces for such industry-leading corporations as US Airways, Wells Fargo, USAA, FedEx, Ford, Walgreens, Telefonica Movistar, Bank of America, Fidelity, US Bank, Geico, Vanguard, Samsung, Virgin, MetroPCS, Sprint and Sempra give him a deep understanding of the user and business challenges of mobile, multilingual and multimodal application design, development and implementation.