Since John McCarthy coined the term “artificial intelligence” in 1955, people have expanded on its original meaning to imagine a human-like intelligence that interacts with its environment, understanding the changing ecosystem and contextual clues which inform its function or purpose.
Three AI characteristics help it fulfill its unique purpose in any given situation: personal, responsive and prescriptive. AI digitally imitates the human mind’s capacity to discern signal from noise based upon the entirety of its inputted (learned) information and external experiences.
The business world has co-opted AI in recent years to represent any number of machine-driven advancements that assist and amplify our business purposes, but as we'll see, not all AI is built the same.
Early Days for AI
Former Director of Stanford Artificial Intelligence Laboratory, Andrew Ng wrote recently in Harvard Business Review: we are very much in the early stages of artificial intelligence.
So given the hype bombarding executives across verticals of all shapes and sizes, how do we make sense of AI's practical applications and impact for business, and, more importantly, how do we evaluate what distinguishes one AI from another?
Our collective interpretation of AI has expanded over the past two years to encompass a number of machine- and learning-driven software. This software uses human-provided construction or “rules” to simplify large data sets into relatively consumable findings. It makes judgments from the data set and responds by either performing an automated action, or maybe suggesting an action to be taken.
Essentially, what we have is advanced statistical modeling, where the AI examines large volume of data points to reach a conclusion, a conclusion which becomes more precise or adapts as more data is introduced.
Is that AI?
On the one hand, a machine did assist its human overlord in computing something he couldn’t manually handle — especially at scale — and then followed up with an action or an insight derived from its rules-driven construct. But is that as uniquely personal, responsive and prescriptive as the term AI should denote?
Maybe, but let’s dig a little deeper into each of these terms to see how AI might impact the creation and management of digital experiences.
The first and probably most important question is:
Is It Personal?
Every person — every digital visitor to a brand, that is — visits your site with an intention. And those intentions are as unique and varied as the individuals themselves.
They are not a segment, they are not a demographic and they are most certainly not a statistically modeled persona. And, while it is a factor, they also are not the same as the last time they encountered you.
Think about retail. The busiest season of the year — the holidays — is all about shopping for others. What you know about the consumer could be highly irrelevant in that instance.
Now let’s add another layer: how visitors connect by channel, device or touchpoint. Whether they come directly to your site, mobile site or they organically search for you or click through an email, every customer journey is crooked, stopping and starting over the course of a conversation.
Enterprises are facing a huge silo challenge, both from a data and a resource perspective.
So while we've all heard the tenets — the world is omnichannel, device-agnostic, etc. — answer me this: how many times do I have to introduce myself when interacting with your brand before I receive personalized attention?
Can Your AI Do the Heavy Lifting?
The point is this: these expressions of intent, touchpoints or channels are only pieces of a complex puzzle that make up a customer.
When a customer walks into a store, he doesn’t expect the salesperson to engage with him based on the fact that he is a middle-aged man who once bought a sweater. Instead, clues about his context, and the salesperson's judgment and experience should inform a coherent conversation that satisfies his intention.
AI should bridge these gaps as well, otherwise we’ll continue offering disjointed and incomplete customer journeys. Yet organizations resist change, putting themselves at risk of death by a thousand segments.
Customers expectations for truly unique and personal experiences are growing and will continue to grow.
So when marketers, site-experience leaders or WCM experts investigate leveraging AI to optimize digital experiences, it’s imperative they evaluate its cognitive ability to automatically adapt to changing customer context. It can’t rely on manually tuned rules or a segment-driven construct.
Instead, it should make the customer experience personal based on pattern-recognition intelligence analyzed from behavior exhibited across a brand’s entire digital presence. It's extremely hard science, because customers don’t necessarily log in every time they visit your site. So being able to match that anonymous visitor with pattern recognition requires a ton of AI processing power.
That doesn’t mean we're off the hook though. Companies like Google are already using AI to set our customer's expectations and if we choose not to leverage it ourselves, we risk obsolescence.
At the end of the day, I will always be “Kevin,” whether I’m known or anonymous. Now it’s up to the brands I frequent to understand that fact throughout my own disjointed journey.
In the next post, we'll dive into AI's ability to be responsive and prescriptive.