Using ReadSpeaker to Create Your Own Digital Voice

When you plan your brand’s voice engagement strategy, what are your priorities?

You could join all the Fortune 50 companies on smart speakers; by 2020, 52% of these enterprises had their own Alexa skills, and 42% offered Google Actions.

You could focus on a branded voice assistant; usage is expected to grow to 8.4 billion AI assistants, globally, by 2024.

You may create interactive voice ads, experiential marketing campaigns, or conversational IVR systems that use AI to give callers the personalized treatment that keeps them coming back. Ideally, your voice strategy will encompass any and all of these voice-first channels.

But to get full value out of any voice initiative, one fact remains: You must sound like yourself. Without a distinct brand voice—and we mean that literally, not in the B-school sense—consumers may not associate your magnificent voice experiences with you alone. When a customer interacts with your brand through Alexa, for example, you don’t want the takeaway to be: “Amazon’s great!”

Make this point to your CEO, though, and you’ll encounter the obvious question: “Okay, but how do I create my digital voice?”

The answer’s simple: Contact ReadSpeaker. We’ll work with you to make your own AI voice, ready for deployment in all your voice-first conversational channels. ReadSpeaker provides white-glove service for companies that need a unique text-to-speech (TTS) voice with market-leading quality. Here’s how it works.

How to Create an AI Voice with ReadSpeaker

Every ReadSpeaker custom voice project starts with a dedicated team of experts, including:

• Your account manager, who acts as a single point of contact between you and everyone else involved in the process. Our account managers ensure open communication and create the blueprints for successful collaboration.
• The ReadSpeaker VoiceLab, a team of computational linguists and AI engineers, who do the hard work of training deep neural networks to produce accurate, listener-pleasing neural TTS voices.
• Ongoing support staff, including speech scientists and technicians, who provide assistance and continual updates for the entire service life of your contract.

This team works hand-in-hand with you to create the ideal vocal representation of your brand. We do this in four steps, outlined below:

1. First, we do a custom text-to-speech voice assessment.

Before we can build the perfect custom TTS voice, we need to understand your business a bit. We start by asking: Who is your audience?

What are their demographics, where do they live, and how do you engage with them outside of voice? Once we have a picture of your buyer persona, we learn all about your brand traits and match them to vocal characteristics that resonate with your customer base.

We also complete a technical assessment. Where do you plan to deploy your TTS voice? The technology we use to deliver TTS to a stand-alone device is different from what we’d use in a mobile app. While you can host a ReadSpeaker TTS solution on a device, on your servers, or in the cloud, it’s helpful to know which of these you need before we begin production. We’ll also determine the geographical regions and markets you occupy, which will inform the languages we need to cover.

Finally, we’ll ask if you have particular industry-specific jargon or terminology. This helps us perfect the scripts the voice talent will record. If you’re in the pharmaceutical industry, for instance, we’ll need to build lots of pharmaceutical names into the script, empowering your TTS voice to pronounce similar terms correctly.

These are the three pillars of a ReadSpeaker assessment: brand identity, technical expectations, and industry-specific language. With these topics covered, we can move on to helping you make your own AI voice.

2. Next, we pass the project to our VoiceLab for production.

This is where we go from the drawing table to an audible custom TTS voice. We start by finding an exceptional voice actor whose vocal qualities echo your brand personality—or, in a voice cloning application, we coach your chosen speaker in the recording booth. (Maybe your brand’s voice is that of a legendary executive, an animated mascot, or a celebrity brand ambassador; it’s all possible through neural TTS.) Voice recordings provide training data that informs our deep neural networks, the AI models that our VoiceLab uses to create warm, lifelike synthetic speech: Remember, every top-quality TTS voice starts with human talent in the recording booth.

We use proprietary recording scripts to obtain every part of speech, from phonemes to vocal inflection. (Alternatively, we could personalize one of our existing 50+ TTS voices—with more in production every year.)

Once we produce the voice, we complete an exhaustive process of quality control and refinement. Then we work with you on further quality assessment, improving prosody, speaking speed, pitch, and pronunciation accuracy until you approve the final product. Soon, we’re ready to put your new brand voice out into the world.

3. Then it’s time to implement your TTS voice in all your chosen applications.

Here’s where your specific TTS use cases come into play. We deploy a consistent brand voice across all your channels: On devices, through a smart speaker app, in your website voicebot, on your IVR system, and more. But each may require a different type of TTS integration. We may use a software development kit (SDK) or an application programming interface (API). You may need deployment through an in-house server, over the cloud, or onboard the device itself.

All these options are available; in fact, given the wide variety of TTS engine technologies at our disposal, ReadSpeaker will ensure your voice can be heard anywhere you need TTS. But even after launch, the relationship continues.

4. After launch, we provide ongoing support.

Your service contract with ReadSpeaker keeps your custom TTS voice functioning perfectly for as long as you need it. This contract includes:

• Automatic technical updates
• Continued account management
• Self-service support through technical documents and product pages
• Voice implementation in new voice channels for permanent scalability
• Access to ReadSpeaker linguists, who will update your pronunciation dictionary to include new proper nouns, fresh industry jargon, and other language changes that would lead to mispronunciation in unsupported TTS applications

This long-term support helps to ensure that your custom TTS voice will remain future-proof, improving with the technology to match your customers’ changing expectations.

Custom Text-to-Speech Voices from ReadSpeaker

When you make your own digital voice with ReadSpeaker, we ensure that this voice remains on-brand—not just in terms of expressing your brand traits precisely, but also in terms of sheer quality. Your voice must reflect the quality of your brand. A do-it-yourself voice cloning app simply can’t deliver that level of excellence. ReadSpeaker will.

In this new voice-first world, these are the table stakes: Your brand is your voice to a consumer responding to an interactive voice ad, or chatting with an in-app voicebot, or calling your conversational IVR system. That voice is the equivalent of a logo, a catchphrase, or any other marker of instant brand identification. ReadSpeaker ensures that, when you create a digital voice, it gets results.