5 Innovative Tools to Build Voice Experiences

Mary Tomasso
Women in Voice
Published in
7 min readDec 17, 2020

--

Listen to the audio version of this article with WIV Soundcloud!

Back in 1966, when the first chatbot ELIZA was created, you had to be a computer scientist to be able to design a chatbot. AI was at its very early stage, and probably nobody would have imagined that in 2020 a linguist would have been able to design conversational AI.

As chatbots and voice assistants became more and more popular, new platforms, software, and tools to build them came to the market.

AI-based and non-AI-based digital assistants

We must make a distinction between tools based on Artificial Intelligence — specifically Machine Learning, Natural Language Understanding (NLU), and Natural Language Processing (NLP) — and platforms that allow us to build digital assistants based on rules.

AI-based virtual assistants will learn from human interaction, which makes them capable of understanding even more complex sentences. Rule-based ones will respond and interact based on rules such as keywords and yes/no.

I turned to Yaki Dunietz , founder and CEO of CoCoHub for an explanation on the difference between both and this was his answer:

“Rule-based is as much AI as machine learning.

For conversational purposes, machine learning is only effective in defining intents, which can easily be achieved with little or no machine learning at all.”

AI-based chat and voicebot platforms

Presently, some of the most popular platforms to build AI-based assistants are Dialogflow by Google, Watson by IBM, Lex by Amazon, Microsoft Bot Framework, and Rasa.

While we are not going into the details of these tools in this post, they are certainly worth mentioning.

Some of these solutions are not affordable, require technical knowledge, or have a steeper learning curve, but they are increasingly widespread, and knowing them might represent a competitive advantage.

Non-AI-based chat and voice bot platforms

These are usually very intuitive, affordable, and easy to use tools. Many allow for the design, prototyping, and deployment of digital assistants, making building a voice experience quite easy.

In some cases, they are a sort of hybrid, and at some point, they rely on AI, but not as much as the above-mentioned tools.

There are so many new platforms being created every day that it’s becoming hard to keep pace, and at a certain point, we need to build our toolkit and make a choice.

These two are my favorite when it comes to designing a voice experience.

Voiceflow

Voiceflow is THE voice prototyping tool. It’s neat, clean, intuitive, and definitely a must-have in the conversation designer’s toolkit.

It is 100% oriented to voice assistants, making building and deploying an Alexa voice skill or Google action a matter of minutes.

You start from a design canvas and build your flow by adding blocks, in a logical sequence starting from the invocation (Home) to the exit.

You’ll find blocks in the left menu within “Steps”. Each block once dragged into the workspace can be fully customized, including choice of voice, pitch, speed, etc.

You could even upload an MP3 file with a custom voice, and build visual screens for Echo Show devices via Alexa Presentation Language (APL).

If you don’t immediately think it’s intuitive, you can navigate through the comprehensive library of tutorials on their Youtube channel. They will walk you through pretty much every step: there’s no excuse for not trying it and mastering it!

You can also choose among a bunch of templates if you are not ready to start from scratch.

Once you link your Alexa Developer and Google Actions Consoles, you can easily deploy to those platforms.

Pricing

These are the current pricing plans. You can start with the free plan and consider upgrading when you have a team because team collaboration between content strategists and developers on Voiceflow can be crucial.

CoCoHub.ai

Maybe I’m biased because I am a CoCoHub certified designer, but I really like the tool, I found the training to be very helpful, and I got to know the platform well enough as to provide an honest and accurate review.

CoCo stands for Conversational Components, and I find that this is a unique approach to bot building, as it allows you to create and deploy a chatbot or voice assistant using pre-built, customizable & tested conversational components that cover a wide range of conversational topics. The pre-built components are in English, but if you take the “name component”, for instance, you can translate it to any language and contribute to the marketplace.

The platform also allows developers who have built valuable components, to become vendors and monetize their components through the marketplace.

This is how the Studio — where you build your bot — looks like.

As you can see, it’s a diagram flow, and each block that you see is a component. Some components are just “Say” which is the bot prompt, and some are just “Navigation” which refers to the expected input from the user.

The little icon on the right bottom corner is a chat box that allows you to immediately test the conversation, both with text and voice.

Once your flow is ready to be deployed, you can easily do so from the “Channels” button, and choose among Alexa, Facebook, phone, embed, and Zoom!

This last feature is actually brand new, and it will make it possible for you to give a face to your digital assistant!

Being able to interact on Zoom with your digital assistant is such a step forward! Think of how strong from a branding standpoint having a digital assistant with a face and body can be!

CoCoHub also has a series of tutorials and documentation to get you started.

You can also find further information on the pricing and I definitely recommend you follow them on LinkedIn and pretty much everywhere, as they are doing great things.

In particular, they have created the first-ever transgender bot that is now becoming a digital influencer: AnnA.

Other tools

Going into the details of more tools would be impossible in a blog post, but I want to mention other tools that for different reasons might be worth testing.

PandoraBots

I haven’t actually tested it, but it’s definitely worth mentioning as they have built one of the most fascinating digital persons, Kuki, considered the best conversational AI. You can find her here and follow her “bot battles”.

The Pandorabots platform is based on an open standard scripting language called Artificial Intelligence Markup Language (AIML) and even if it doesn’t seem that intuitive to me (my bad), you can access comprehensive documentation here and learn how to build your digital assistant.

The Voice Designer

This is a WordPress plugin that lets you build a voice-based version of your website or blog and deploy it directly to Alexa and Google Assistant.

From the backend of your WordPress website, you can create a voice bot, customize it, and build it with blocks, similar to VoiceFlow.

Engage

As the name suggests, this platform’s aim is to help companies and brands engage with their audience by voice.

You start by entering your brand’s name

You choose a template:

Just some small customization and the voice “engagement” is ready to be deployed, but before, you can test it:

Unfortunately, when I tried to test this flow on Alexa with my name, I had some problems with my name not being recognized, maybe my pronunciation wasn’t recognized as my name is Italian and so am I.

At a certain point, I thought I had broken Alexa: there was an infinite Loop of “Sorry, I could not find a voice engagement called Mary Tomasso, please try again.”

I finally changed the name to something easier to pronounce, and I got to experience the whole flow.

I think it might be a good solution — perhaps too pricey — for large companies to easily engage with their audience.

Conclusion

When it comes to conversational AI, it’s not all about the tool, there are so many human skills, experiences, and expertise involved in the process.

However, the right tool will be like your magic wand, so choose wisely, based on your needs, your budget, and — a very important requirement for me — the availability and support you might get from the chosen platform.

Do you have already a chosen platform? Let us know in the comments!

--

--