Device Type: desktop
Skip to Main Content Skip to Main Content

Artificial Intelligence and Customer Experience: The Real Deal

This article was published on July 13, 2021

You’d be forgiven for thinking that the rest of the world is part of some special artificial intelligence (AI) club, while you’re left wondering where to get started. So, here’s the big secret about AI and customer experience (CX): almost everyone is in the same position as you. It can be hard to see past the hype, but there are genuinely useful ways that AI can help your CX today.

Illustration of robot with headset

Now for another secret. AI is both less and more impressive than you might expect. If you’re using the right tools, then AI-powered tools can slot into your workflows, but in a limited role. With the right preparation, though, they can help take the load off human teams. In particular, there are three areas in which AI can have a meaningful impact on CX

  • uncovering deeper insights into what customers need

  • shielding your human CX staff from simple queries

  • giving both your customers and your staff the information they need before they even know they need it

There are already mature products in each of these categories that you can integrate into your existing CX workflows. And while none of them will be the all-singing, all-dancing human replacement that Hollywood and, let’s face it, some pundits would have you believe, each of these artificial intelligence solutions will add immense value to your customer experience.

AI Is Software That Writes Its Own Rules

Let’s take a moment to think about what we mean by AI.

Most software is anything but smart. At some point, a human had to tell it precisely what to do in every possible scenario. A chatbot programmed in that way can handle only those situations anticipated by its creators. When it encounters something for which it isn’t programmed, the best it can do is connect the customer to a human operator.

Most AI used in customer experience is only a little more advanced than that.  Instead of telling today’s AI-powered CX tools exactly what to do, you can give them some initial guidance and then let them learn from experience.

In effect, AI-powered software writes its own rules. That opens some remarkable possibilities. Imagine a chat bot that recognizes which responses tend to lead to better outcomes and then adjusts how it deals with customer queries accordingly.

Uncovering Customer Needs

For more than a decade, companies have gone all-in on data. Big data, medium data, small data: We’ve been collecting it, analyzing it, and acting on it.

If you want to understand what the next decade holds, then think of that data as seeds sown in a field. Today’s data science techniques have made companies more efficient, unlocked previously hidden insights, and helped deliver better customer service. But it’s as though we’ve been using scythes and manual labour to harvest the crops.

AI analysis of all that data is going to have the same impact in the coming decade as the combine harvester did for agriculture in the mid-20th century.

In customer experience, we’ve become adept at recording customer behavior. We track clicks on websites, we record customer calls, we use beacons to map people’s journey through physical stores. So far, analyzing that data has required human intervention. As an extreme example, think about reading through hours of call and chat logs. Most analysis, though, isn’t quite as manual. It does, though, require that human data scientists know up front more or less what they expect to find.

What if we could take the rich seam of data from calls, emails, text messages, and chatbot conversations and automatically find insights that improve how we do business? Today, it would take a human team years to sift through every customer interaction. And by the time they were done, they’d be in no fit state to make an analysis.

There are, though, tools that use AI to unlock the insights held within all that data. Let’s unpack what that means.

With so much data, it’s possible to find patterns of activity that correlate strongly with certain outcomes. Those endless hours of recordings and gigabytes of text-based communication seem impenetrable to a human. An AI tool, though, can look for patterns in that great sea of data and make connections between certain patterns—such as a turn of phrase, the time a call was placed, or the weather—and subsequent outcomes.

Such historic data can show that certain phrases mean that a customer is very unhappy even before they’ve explicitly said so. Or that a transaction is fraudulent.

For the AI, it’s just a matter of looking for patterns between what was said and what eventually happened. Such tools come primed with some patterns but, crucially, they can learn new ones. That means they can, amongst many other things:

  • monitor sentiment, enabling managers to see which staff and which actions result in better outcomes

  • pinpoint the techniques that help top sales performers to sell more

  • detect key phrases and ensure service quality levels are met

Using your existing data to get a better understanding of customer needs and your CX performance is one way that you can use AI right now. So, what about using AI to deal directly with customers?

Giving People Time Back

Let’s not beat about the bush. Right now you’re probably wondering when we’re going to get to chatbots.

Depending on who you listen to, it can often seem that there isn’t a problem that chatbots can’t solve. The reality is a little more nuanced.

Today, there are two types of chatbot: those that use AI and those that don’t. You might have interacted with the second type and come away frustrated by its lack of understanding and narrow range of responses. Such chatbots look out for fixed keywords and phrases, then offer a pre-canned response in return.

Now, we’ve all seen the stats that show that high-quality customer service is a key factor in avoiding churn. Choosing the right type of chatbot can help you deliver better customer service in three ways:

  • first-line queries can be handled instantly

  • human agents have more time to look after those queries that must have human attention

  • customers get a conversational way to self-serve—and customers increasingly love self-service

So, how do you create an AI-powered chatbot without enlisting a team of AI specialists? Vonage partner OneReach provides a no-code platform that lets you create AI-powered bots without needing to be an AI expert or even a developer. What matters is that you understand the needs of your customers.

Using OneReach, you specify the channels through which the bot should communicate—such as SMS, Facebook Messenger, or voice—and then seed the bot with the initial knowledge it needs to get started. From there, though, the OneReach platform uses AI to learn how customers interact with your bot, and it grows in its ability to serve them.

The conversational aspect of the bot is just the user interface, though. Through OneReach’s drag and drop system, you specify how it integrates with your existing systems, such as Salesforce or Zendesk.

Those integrations and the AI-powered interface mean that the bots you create function as another direct interface to your systems. So, you can build conversational interfaces to functionality that would otherwise be accessed through a mobile app or web dashboard. That way, customers can use WhatsApp to get instant updates on the progress of an order, book a restaurant table using a voicebot over the phone, and more.

At that stage, your chatbot starts to become something more. It becomes a virtual assistant.

Augmenting Humans

Virtual assistants are already a part of our lives. Thirty one percent of US homes have a smart speaker, and billions more around the globe have access to virtual assistants on their smartphones. Whereas chatbots tend to be singular in their purpose, virtual assistants bring together multiple services to make life easier for people.

When it comes to customer experience, virtual assistants have a role to play on both sides of the conversation. For customers, a sophisticated virtual assistant anticipates their needs rather than only responding to requests in the way of a chatbot.

Let’s take a simple example. With appropriate permission, a bank’s virtual assistant could analyze a customer’s account. Based on previous transaction history across thousands of customers, it might recognise that people whose accounts suddenly start showing purchases at baby product stores then often go on to open a college savings fund. The virtual assistant could use that pattern matching to offer a college savings account via SMS to people whom it identifies as prospective parents.

On the agent side, virtual assistants can anticipate customer needs by listening in to a conversation and ensuring that the agent has the information they need before they even know they need it. IBM’s Watson platform, in partnership with the Vonage API platform, can observe voice calls and text-based conversations in real-time and then present the human agent with what they’re likely to need.

AI Is Part of CX Today

Platforms such as Watson show that artificial intelligence is here right now. It’s no longer only for sci-fi. Instead, it is playing a part in the customer experience delivered by companies in just about every industry.

Working with Vonage APIs and their partners, you can integrate such AI solutions into your existing legacy contact center platform or into your cloud-based platform. While it’s easy to get distracted by sci-fi expectations of AI, the reality is still pretty exciting. More than ever before, software platforms can help us deliver higher quality customer experiences and reduce the burden on our front-line customer experience colleagues.


Vonage Staff

Deskphone with Vonage logo

Talk to an expert.

AU free phone number: 1 800 239 458