Published: July 08, 2020 | Comments
Every day, artificial intelligence helps us accomplish more than we ever thought possible. One such example is the way we access information. We expect immediate answers to our questions, and AI makes that possible. Likewise, we see that the obvious application of AI is scale; we reach far more customers with a simulated human experience than with staff alone.
However, we shouldn’t think of AI as a one-way street where information is passed from an algorithm to an end-user. Instead, we should recognize that valuable information also flows back from the end-user. If we pay attention to what, when, and how people are communicating with AI, we can make better informed decisions that can enhance the customer experience.
Sophisticated AI systems rely on Natural Language Processing (NLP), which essentially matches an input against a stored label. So, if a user asks “what’s your name?” and the AI has knowledge stored against labels such as “what-is-your” and “name”, it will return the information associated with that specific pair of labels. We can then measure the frequency with which users ask about specific topics to identify which subjects are actually most important to our users.
Similarly, if the AI is missing information for given labels, we can gain insight into questions users are asking that we might not have addressed in our resources. Understanding the specific language people use to ask questions is also valuable, as this reflects the way they will search for information on your website. An accurate representation of what our customers want to find on our website can help us create more intuitive, informative web content. Evaluating the “what” of conversational AI analytics makes it possible to produce the customer experience more responsive to individual customer needs.
Another valuable insight gleaned from transaction data is when customers tend to seek information. Traditional models rely on an assumption of customer preferences, since service centers might have operational hours that will otherwise influence behavior. For example, a center that is open from 7:00am to 6:00pm might experience peak volumes between 12:00 pm and 1:00 pm. This may certainly represent the most convenient time for people during those specific hours. However, if AI is in place to assist users 24/7, we may find that users are most active between 7:00 pm and 8:00 pm, guiding us to shift our operational hours to accommodate that preference. Likewise, transaction data provides a more accurate representation of when users prefer to communicate with us, and we can leverage that data to improve their experience.
We can also make improvements by analyzing how users interact with us. AI systems support nearly unlimited deployments; they can live on your phone system, your website, your SMS texting platform, your social media pages, and even our virtual assistants like Amazon’s Alexa. Conversational AI analytics offer a glimpse into which channel is most important to users. With that information, we can make informed decisions about where to invest additional resources, and which channels to promote to our customers. Additionally, you might find that it’s best to highlight certain information on your social media pages, and other information on your website. If we understand how customers seek information, we can tailor our delivery systems to reflect those preferences.
Industries tend to have unique qualities, and it’s unlikely that these examples offer a comprehensive summary of how AI informs organizations. The important thing is that we pay attention to what the data is telling us. When we do, the decisions we make to shape the customer experience are better informed, and more likely to add value for our stakeholders.