The Synthetic Voice as a Consumer Engagement Tool

Inventors have been trying to imitate the human voice since well before the dawn of digital computing. As far back as the late 1700s, Hungarian scholar Wolfgang von Kempelen built a mechanical speaking machine that produced an honest-to-goodness early synthetic voice. The machine’s bellows, reeds, and rubber articulation cup may have sounded more like a dyspeptic cow than Siri or Alexa, but it was a start.

Not to discount the genius of the past, but von Kempelen’s speaking machine wouldn’t help much with a modern smart speaker app. Today’s synthetic voices are powerful precisely because they imitate human speech at scale. A digital synthetic voice is a scalable imitation of human speech, and this expandable deployment is a primary benefit for businesses: It offers one consistent brand voice for all your audio channels.

To understand synthetic speech, start with the modifier, synthetic, as in synthetic fabrics or, to crib from pharmacology, synthetic molecules. They’re manufactured versions of natural things. The same is true of a synthetic voice—and, as with other synthetics, the manufacturing process makes all the difference in terms of quality.

How to Develop a Personal Synthetic Voice for Your Brand (And Why It Matters)

The most common type of synthetic speech today is text-to-speech, or TTS. This technology starts with human voice recordings. Then, in the sort of neural TTS we create at ReadSpeaker, engineers use those voice recordings to train a deep neural network (DNN) model, which relies on advanced machine learning to predict accurate pronunciation for any text. The trained DNN model translates the written word into spoken language that sounds remarkably like the source speaker (see sidebar).

If you need an AI robot voice to power your voicebots, owned personal assistants, smart speaker apps, or any other form of conversational AI, ReadSpeaker AI will work with you to develop a future-proof synthetic voice that represents your brand traits—and provides instant brand recognition across all your platforms.

We start by listing the characteristics that define your brand—is it rugged and outdoorsy or coolly chic, for instance? Either way, we’ll find a voice actor that expresses those traits through voice tone and speaking style. Then we’ll develop a synthetic voice that carries that natural brand representation consistently across voice channels. This operates in the audio space the same way your logo does in graphics: It’s a crucial differentiator in a crowded marketplace.

Benefits of Synthetic Speech Throughout the Sales Funnel

Many innovation teams make customer service a priority; that’s often what brings organizations to the concept of a synthetic voice in the first place, and for good reason. Without a custom TTS voice, you can’t provide a recognizable brand experience in voice-driven engagement platforms such as:

• Contact center (CC) solutions like conversational interactive voice response (IVR) systems and AI-driven intelligent virtual agents (IVA)
• Voicebots and branded personal assistants on websites, mobile apps, and smart home devices
• Automated customer self-service apps on smart speakers and other voice assistant platforms

But the benefits don’t stop at customer service. In fact, voice-first digital engagement can help with the entire sales cycle. Adding TTS to your website or mobile app improves accessibility for people with vision impairments, visitors with reading disorders, second language-learners, multi-taskers, and users who simply prefer audio content to the written word—all of which expands your audience considerably.

Conversational AI can also help convert those contacts into leads. A voicebot that answers top-of-funnel questions may forward data to the sales team, providing hyper-personalized information about what would-be consumers want from your brand.

Voice-first conversational AI can also help with the educational efforts that both grow your audience and encourage ongoing engagement. We’re talking, of course, about content. Like any brand, you push out written content. You create video content. Voice-first channels provide a third universe of content-based engagement, and with conversational AI, the information flows both ways.

At the most basic level, you can incorporate a synthetic voice to transform your blog into a podcast, reaching a broader audience and getting more value from existing assets. Or you can go fully conversational. Imagine a bank mobile app with a voicebot that provides personalized advice on budgeting or investing. These sorts of educational resources go beyond the blog, expanding audiences and improving brand engagement to keep your customer relationships going and growing. But why use a synthetic voice when you could simply record a voice actor? A few reasons.

Neural Synthetic Voice Vs. Traditional Voice Recordings

Before TTS became broadly available, the only way to scale a voice was through recording, which is just how brands did it when radio and broadcast television were the dominant customer engagement media. There’s still an important place for the recorded human voice, but some use cases absolutely require TTS.

A synthetic voice can provide functionality that voice recordings cannot, as in a conversational AI system that uses natural language generation (NLG), or AI writing, to compose helpful responses based on user prompts. With voice recordings alone, audio responses are preplanned, canned, and static. With conversational AI and TTS, the bot can say anything the NLG module comes up with—and only a synthetic voice makes this possible.

Similarly, TTS is more scalable than voice recordings. Need a last-minute script change? Just type it into your TTS engine; no need to book the original talent for a make-up recording session, assuming they’re still available and interested. But these are generally short-form applications. For longer speech that has no need for dynamic content generation, many still choose voice recordings over TTS. And that’s great! There’s room for both in the developing Internet of Voice—and a custom synthetic voice can help you survive and thrive in the voice-first marketplace.

The Future of the Personal Synthetic Voice

In the past, Wolfgang von Kempelen’s mechanical experiments were the best synthetic speech available. Today, we have neural text to speech. So what’s next? We’ll find out soon enough, as TTS technology is charging ahead at a tremendous pace. But a few trends are already clear. For business users, synthetic speech is rapidly going from nice to have to necessary, a progression driven in large part by the 2020 pandemic’s sudden shift to remote communications between brands and consumers.

Advancing deep neural networks and voice-text markup languages are also ushering in an era of emotionally expressive TTS. Where early TTS sounded like a human voice-robot hybrid, situational inflection is now creating friendly, even comforting synthetic speech—a capability that results in a much better customer experience with every engagement.