Getting Started: Using AI to Make Connections With Customers
/
We set the stage for a series of posts (#2 here) looking at the role of artificial intelligence (AI) and customer interactions as they relate to product and service innovation. In this first post, we offer some fundamentals for customer-facing intelligent virtual assistants. Next up, we'll link these digital assistants to product and service development teams. On the horizon, look forward to considering how human representatives engage with their digital assistants to augment their human capabilities.
Thanks to Utpal Mangla, VP and Senior Partner, IBM, for joining me in this post. Under Utpal's leadership, his team recently achieved its mission of making "Watson AI Impact 1 Billion Consumers." Utpal and I are officers within the ISSIP (International Society for Service Innovation Professionals) organization.
Artificial intelligence may usher in a golden age of customer contributions to product and service innovations. Customer service interactions offer opportunities during the interaction itself and then, in a more macro sense, from the organization's ability to glean insights from trends across interactions. However, there is never a silver bullet -- we must introduce new technology tools alongside training for customer service and product development talent and new techniques that keep the knowledge flowing from customers to people in design roles.
Some Opportunities Offered by Digital Assistants
Intelligent virtual assistants (also known as chatbots/virtual assistants/digital assistants) are tireless at documentation and remain calm under fire. In some cases, people are more willing to disclose information to a chatbot. These attributes signal wide-ranging opportunities, especially in customer-facing roles.
We've seen that combining human customer service representatives with intelligence virtual assistants can double productivity with a 50% reduction in cost. Additionally, while customer care employees are often bound by scripts in their interactions with customers, artificial intelligence systems are allowed to learn and change responses based on broad access to customer interactions and outcomes.
Fundamentals for Customer-Facing Digital Assistants
The fundamentals of customer centricity remain unchanged while working with artificial intelligence. Performing customer segmentation to identify the target audience, understanding the customer needs, and tailoring responses to match those customer's needs are all critical for the success of digital assistants, and all of us.
Based on over two decades of experience working with customer interaction systems, seven years experience building AI virtual assistants, and continued work in the development of these intelligent digital applications, we offer a checklist of 'best practices' to maximize the success and productivity of digital assistants from your customers' perspective -- this is foundational to the digital assistant being a resource to your product/service development team (Part 2 of this topic, coming soon).
Ease of Use: Seamless usability is critical for success. Keep it simple. Successful digital assistants are easy to work with, have reduced clicks, use voice interaction, and are concise in communication with the customer. For an example, see Vodafone's TOBi.
Datasets & Training: Data is essential to developing a truly conversational digital assistant. Digital assistants, like all of us, learn and become intelligent over time. As we deploy an initial set of "intents," it's important to ensure that the intents have been fully tested across various data sets. A digital assistant is only as good as the data it is trained against. Here's an example from Stanford University's Open Virtual Assistant Lab so you can see the issues in context.
What is an intent? An intent is defined as a particular activity or objective that the customer has in mind when interacting with chatbots. Examples of intent include: Paying a Bill; Check Balance in Checking account; Return Products
Land & Expand Strategy: Start small and scale over a period of time. Start with a limited set of intents. Follow a phased approach to adding more features & functionalities. A small set of robust intents functioning accurately is more effective than a large set of intents that don't solve customers' needs. You want the customer to start liking the digital assistant and asking for more of it. Mostly for fun.
Context-Based Workflows: As virtual assistants take customers to next steps, it is important that the workflow successfully completes the specific activity or answers all queries tied to that transaction. Workflow can be an "intent" or a "combination of intents." When customers start doing more complex transactions, it is a reflection that they like the digital assistant and want to take greater advantage of new features & functionalities. The Zappos Customer Loyalty Team is amazing. If you're building a customer-facing digital assistant, the Zappos approach may be the holy grail for the combination of your digital and human (augmented) customer team.
Omnichannel experience: Successful design strategies provide a seamless omnichannel experience (mobile, voice chat, phone, social media, etc.), regardless of which platform they use. Virtual assistants integrated with the channels can help solve this problem.
Measurements: Measuring effectiveness is critical to success. Measurement should be done from both customer and business angles since they are closely interrelated. Some of the commonly used key performance indicators include Net Promoter Score, Customer churn, Customer effort score, First call resolution, Average handling time, Reduced cost per transaction. Note that Zappos uses Net Promoter Score as well as some other metrics that will stretch your thinking of your digital assistant's goals.
Next Steps Toward Innovation
You may well be on your intelligent virtual assistant journey. "As of 2020, 54 percent of customers have daily AI-enabled interactions with organizations, including using technologies such as chatbots, digital assistants, facial recognition or biometric scanners." Hopefully, your organization is aligned with the fundamentals above. Our next step is to extend the value of these AI interactions to your product and service innovation teams. Check back soon!
