Of all technologies that impact customer experience (CX), few have as immediate or obvious an effect as artificial intelligence (AI).

AI is a natural fit for customer-facing applications because it encourages a more conversational approach to interactions and is easy for customers to use. Just about everybody reading this has encountered some form of AI, usually in the form of chatbots, speech recognition, or recommendation engines. However, most have encountered more subtle forms of AI such as contact center call and ticket routing, image processing, and analytical tools used by businesses as well.

The market for AI applications is young and growing fast. Current market size estimates range from $10–15 billion, with the market expected to grow to over $100 billion by 2025. It’s estimated that AI has created $2–3 trillion in value to date. Clearly, somebody—many somebodies—thinks this technology is a good idea, but why?

What is AI?

Douglas Adams’ The Hitchhiker’s Guide to the Galaxy described artificial intelligence-driven robots as, “Your plastic pal who’s fun to be with.” More seriously, AI is how we describe machines that mimic cognitive functions of living beings. This includes the ability to make inferences and fill gaps, respond to body language, make decisions based on incomplete information, and learn from experience.

Describing its capabilities, components, and challenges would be enough for a weighty book all by itself, but the practical development is that AI enables machines to operate autonomously, which allows humans to act naturally when working with them to achieve expected outcomes. AI is learning to drive cars, identify individual faces in a crowd, and help an agitated airline passenger get an alternate flight. The best AI can interpret and respond to human emotional states, so maybe Adams wasn’t that far off. When’s the last time Alexa laughed at you for no reason?

For customers 

Some of you have used Twitter to order a pizza. That’s okay, I’m not here to judge you. Presumably, you knew you were interacting with a chatbot, and a live person wasn’t taking your order. Whether it’s ordering a pizza, updating your business cards, or getting an insurance quote, bots—AI-driven virtual agents that engage with customers and perform business processes—are some of the tools businesses use to communicate with customers via social media with minimum delay or misunderstanding.

Some bots take orders or collect poll responses, while others connect you with the person who can help you best based on its understanding of your request. Both use natural language processing (NLP) software to analyze your words and their context to determine the best course of action.

Let’s go back to pizza (yay!) and work through some hypotheticals. If you had used the bot before, you could probably say, “Hey PizzaBot, give me my usual and send it to my house,” and you’d receive your most common order (or one you had predefined as your “usual” via an app) at your home address. What if you had no address defined as home, though? The bot might ask you for an address, or confirm that a previously used address was the correct one. It might just assume that the only address it has in your record is home, depending on how the bot has been instructed. What it won’t do is fail outright because you gave it an invalid request—at the worst, it will connect you to a live agent to complete the transaction.

The same is true of a nonexistent “usual” order. It might clarify, confirm, or assume based on the available data, and send you to a human if it can’t resolve the problem. A more advanced AI with the proper integrations could guess your usual order based on the most popular orders in your area, modified by known demographic data on you, the time of year, or a number of other factors. It’s worth noting that this would not necessarily be a better choice, just one that considered more variables. It’s possible to be too smart for your own good, this is true for machines as well.

For business users

Artificial intelligence doesn’t have to be visible to the customer to play a role in customer experience. One of the first business uses of AI was to enhance the value of predictive analytics software and similar business intelligence (BI) tools.. Today, AI in the BI space is often referred to as insight engine software, prized for its ability to sift structured and unstructured data and make unapparent connections.

While big data tools might affect how a business approaches changing market conditions, it might not have a huge bearing on individual cases. The business application most likely to impact daily customer experience in a subvert way is the AI sales assistant. This technology can automate a variety of tasks, including lead qualification, data entry, and scheduling, freeing up salespeoples’ time. That’s a net positive for the customer already, but some AI sales assistants provide real-time analysis to the salesperson during a call, extracting relevant product info or customer history, evaluating the probability of a win based on shifts in the conversation, and more. 

Potential drawbacks

So, if AI makes tasks easier for buyers and sellers, what’s the downside? It might not be the gradual replacement of humans by machines in the job market, though that’s something to watch out for. Rather, it’s the effect that instant gratification could have on the human psyche. To quote AI industry trendwatcher Jarno Duursma, “Discomfort is the gateway to reflection and growth; AI insulates us from discomfort.” Without bad experiences, he suggests, we do not experience the discomfort of frustration, waiting, or denial, and thereby inhibit our personal growth.

As a result, we create high demand for perfection in our daily lives—perfection that isn’t always attainable with humans or machines. This expectation of perfection inevitably reduces our satisfaction over time, so something considered good last week isn’t good anymore because we always want something better.

Reliance on AI can lead to the erosion of skills. After all, why learn how to forecast sales or listen to the undertones of a person’s voice if a machine can do it for you? There’s a reason animators still learn to draw moving images on consecutive sheets of paper, even though the vast majority of animation is done on computers with Wacom drawing tablets: Mastery of the basics supports proficiency. To grasp AI’s value to CX (and the value of technology to CX), it’s not enough to understand what the AI is doing but why it’s doing it. These are not disconnected processes, but parts of an overarching experience of what it’s like to do business with your brand. Each step matters.

The future

First, the good news: AI cannot fully replace humans, nor do we have to worry about welcoming our new robotic overlords. General AI, the term used for artificial intelligence that is not focused on specific competencies and is more like sapient thought, is a long way off if it’s even possible. We’ve figured out how to put a small set of skills and behaviors into a machine, but imagine putting everything you know—learning that began the moment you were born and includes conscious, subconscious, and unconscious thoughts—into a computer brain. We don’t even fully understand how we learn or remember, or the extent of our retained knowledge, so how are we going to make a robot do all of it?

That said, AI will increase its capabilities and its use cases will expand. Some jobs may disappear as we turn more tasks over to automation. Even so, somebody has to program those machines, manage them, and develop new ones. And there’s a lesson to be learned from applications like AI sales assistants: they don’t do the job, they support the person doing the job. It is likely that the development of AI will be of the invisible, behind the scenes variety, performing tasks alongside us in response to our needs. 

The future is never certain, but it should at least be interesting.