Do customers really want voice chatbots?

Chatbots have been through a lot. They’ve seen a hype cycle, a growth in AI-powered ability, and they’ve shed their ‘emerging’ label to become a growing customer service staple.

Now, some chatbots represent another new evolution of the technology. Namely, we’re seeing the emergence of voice chatbots.

So, what are voice chatbots, and do customers really want them? Here’s a closer look at voice-powered chatbots and what they bring to the chatbot service table.


What are voice chatbots?

Voice chatbots are chatbots that can communicate using vocal input and output. You can talk to the bot out loud, like you would a person. In turn, it will respond with a voice of its own.

Chatbots of any kind are software programs that interact via conversation. Usually, this works through text-based input and output. So, you type a message to the bot, and the bot sends a message back with relevant information or follow up questions.

Voice chatbots do the same thing, but through vocal input and output. They listen to you talk to them, and then reply verbally.


Types of voice chatbots

There are two types of voice chatbot.

The first is a hybrid bot. These are text-based bots with an added layer of voice recognition over the top. This means that they can support both text-based and voice-based conversations. So, users can choose how they interact with the chatbot.

The second type is the bots that only operate through vocal communication. A good example is your smart home assistants (i.e. Amazon Alexa or Google Assistant.) They exist in voice-controlled devices. Often this type of bot acts more like a general assistant than a specialist bot. But, as they grow in popularity, they’re starting to make their way into more specialised customer service.

For instance, McDonald’s using voice chatbots to manage their drive-thru orders.


How do they work?

Voice chatbots use speech recognition technology which includes natural language processing, or NLP. These are subsections of artificial intelligence concerned with how machines and humans interact.

Speech recognition allows voice chatbots to recognise when they’re being spoken to. NLP, meanwhile, allows them to understand language as it is spoken naturally. So, you don’t have to talk like a robot to interact with voice chatbots.

Voice chatbots are also more often intent-based, rather than flow-based. This means that they work by looking for the intention behind your input. This is because it’s harder to make a flow-based voice chatbot without the conversation sounding too robotic.


So, do customers want them?

Voice chatbots are a technology that we’re growing used to. But will customers want them as part of their service options?

Already, chatbots are proving a popular channel for customer service, with 67% of customers worldwide using a chatbot for customer support in 2018. Plus, voice assistants are popular when it comes to making purchases and dealing with businesses. On average, 80% of those that have shopped using their voice assistant were satisfied.

Both these statistics suggest that yes, customers do want voice chatbots. It’s easy to see why you might sing their praises. After all, 90% of human interactions use voice. So, it makes sense that there’s space for voice chatbots in customer service.

But do customers really want voice chatbots? Another good way to look at this question is to look at the pros and cons of voice-based bots.


What do they bring?

The key benefit of a voice chatbot in most cases in that they represent a hands-free automated service where this hasn’t been achieved before. For customers wanting to perform a simple process – making a booking, placing a repeat order – voice chatbots allow them to do so without lifting a finger.

Plus, they use a technology that many are familiar with. We have speech recognition in our homes and on our phones; we’re familiar with the concept of getting what we want via voice command.

But what else do voice chatbots brings to the customer experience? The answer is much the same as with other bots, but with a twist. For example, their boost to accessibility.

Chatbots are an ‘always on’ contact channel. They don’t need to sleep or take holidays, and they can help any number of people at once. So, customers can always connect with a chatbot. Voice chatbots combine this with the convenience of vocal conversation. So, customers might not be able to phone you at 3 am, but they can still talk to your chatbot representative.


On the other hand…

But there’s a caveat to customers wanting voice chatbots: they can’t replace your human team, and they can’t go it alone. The danger of voice chatbots is the temptation to use them to replace human team members. The issue with this is that without a human to back them up, chatbots have no fallback when they fail.

Take the McDonald’s drive-thru example. What happens if the voice chatbot misunderstands or malfunctions, and there’s no human to step in? It’s an invitation to frustration.

Or take robocalls that see bots making outbound calls and speaking recorded text out loud. It’s notorious how hated these systems are – so who would want this mirrored in a customer support exchange?

It’s important to remember context. A voice chatbot at a drive-thru where no small-talk is required might work well. A voice chatbot manning a customer service telephone line is unlikely to offer a great experience.

Different customers and different support needs call for different service options. Some customers might be happy to interact with a voice chatbot. Others might feel uncomfortable with it.


What are the concerns?

Privacy is a major concern when it comes to voice chatbots — 28% worry about the privacy of their sensitive data and the security of using voice assistants. Meanwhile, 45% say they don’t feel comfortable sending payment through voice assistants. (A common form of voice chatbots.) So, for functions like setting up card payments and sharing sensitive information, it’s likely that fewer customers want voice chatbots.

Voice chatbots are also reminiscent of interactive voice response, or IVR. (Those automated phone trees which are almost universally hated.) This isn’t to say that voice chatbots are the same as IVR. In general, they’re more advanced and so hold more natural feeling conversations. But the moment a voice chatbot fails, they’ll likely come across as frustrating as IVR.

As with any chatbot, voice chatbots create bad experiences when they don’t work. Beyond a frustrating experience, though, this can also impact trust. If a voice chatbot doesn’t understand a command, how can customers trust the answers it provides?


Another choice

Do customers want voice chatbots? It depends on the customer and the specific service situation.

Many are singing the praises of voice-based bots. And, while they’re set to grow in popularity, customers will likely only want them as another way to interact. Not as a replacement for your human team.

Most customers like to have a choice, and there’s definitely space for voice chatbots in customer service. But they aren’t the only channel customers want.


Useful links

Why the chatbot hype fell short

The problem with generalist chatbots

What is an intent-based chatbot?