Home / DeveloperSection / Articles / Voice Commerce: the guide to optimizing your voice journey using UX

Voice Commerce: the guide to optimizing your voice journey using UX

Voice Commerce: the guide to optimizing your voice journey using UX

Talkit AI 877 27-Jul-2021

If you have a smartphone or tablet, the likelihood that you've ever used voice control to search or ask your device to do you a favor is high.

Although France is lagging in using voice assistants compared to the rest of the world, this trend is not weakening, quite the contrary.

However, the opportunities inherent in the voice sector remain timidly exploited.

However, these interfaces where everything is still to be explored are among the most promising sources of growth for the years to come. 
In this guide to Voice UX Design, we'll find out what Voice Commerce is and how UX adapts to these new realities to deliver fluid and engaging voice experiences.

Voice search, interfaces, and voice assistants: what are we talking about exactly? 
Before getting to the heart of the matter, it seems essential to us to distinguish between three fundamentals of the vocal universe:

  • Voice search,
  • Virtual assistants,
  • Voice interfaces.

Voice search, first of all, is to use the voice instead of the keyboard to search on a website or an application.

In strong growth in recent years, voice search is a technology closely linked to the boom in mobile use.

Present on almost all smartphones worldwide, it provides a pragmatic response to the constraints of web browsing in public transport, in the street, or stores, for example.

It also invests the domestic space through the intermediary of intelligent personal assistants, making it possible to obtain information without going through a physical medium.

Personal assistants (or voice assistants) are software applications that use speech recognition to provide answers to user searches. These responses are reproduced using text-to-speech technologies.

The services offered by voice assistants are very broad, ranging from the day's weather to creating a shopping list, launching a playlist, or purchasing a product.

In 2021, four players are competing for the (juicy) virtual assistant market:

  • Apple with Siri since 2011, available on all Apple brand devices,
  • Amazon with Alexa since 2014, whose success is largely attributable to the ability to buy directly on the site from the assistant as well as to Alexa Skills,
  • Microsoft with Cortana since 2014, available on machines equipped with Windows 10,
  • Google with Google Assistant since 2016, deployed on all Android devices as well as on Google Home, for example.

Finally, the voice interface (or VUI) makes the interaction between a user and a voice assistant possible.
Therefore, a good voice interface must be sufficiently clear and educational to suit newcomers and sufficiently effective and efficient to serve professional use, for example.

So you see us coming. The design of a voice interface is CRUCIAL.

Voice search: where are we in 2021?


Far from being a novelty, voice search is, on the contrary, widely used in France and internationally.

Thus and according to Google, nearly 30% of the world population would already use voice search in 2021.

There were also nearly 3.2 million voice speaker users in France in 2019.

If the followers of voice search like to inquire about the weather, the consumer is not left since transactions by voice, called Voice Commerce, could represent 3.5 billion pounds in 2022, nothing that 'in Great Britain

What is Voice Commerce?


Voice Commerce qualifies all transactions carried out by voice command from a telephone, a computer, or even a connected speaker.

It encompasses the entire purchasing journey, from finding the product or service to settling the transaction.

Easy to learn and fluid, voice shopping applications offer a new shopping experience for their users.

They are all the more attractive since the start of the health crisis with the explosion in the adoption of contactless payment, which saves time in the store and respects physical distancing.

Unsurprisingly, this trade is already of interest to brands and particularly to mass distribution players.

By offering its voice commerce experience in 2017, Monoprix was a pioneer. Intermarché quickly followed suit in 2019, followed by Carrefour more recently.
The application offered by the latter has also been particularly distinguished with features of product selection, modification of the basket, and payment, controlled entirely by voice.

But Voice Commerce is not just a simple adaptation of its site for audio transcription. Perhaps more than any other interface, voice requires full integration of the end-user into their experience design.

What is Voice UX, and why is it the discipline of tomorrow?


Voice UX is the experience a user experience when using a voice application.
Like the classic UX, the Voice UX strives to offer a fluid and logical journey and a fine and rapid understanding of search terms to limit the user's frustrations.
The professionals of the vocal experience are the Voice UX Designers. They are responsible for creating rich dialogues relevant to the user's culture, with the right words, the right intonations, the right flow...

Therefore, the voice interface design requires both a technical platform (ergonomics, development of concepts ...) and a more emotional base (creativity, empathy for the needs and difficulties of the user ...).

Billions of smartphones are now equipped with microphones promising a very bright future for this new profession. Still, in its infancy, Voice UX is a boulevard of creative opportunities for journeys and experiences that have yet to be invented.

Voice UX Design: designing an engaging voice interface in 8 steps

No vocal experience is alike. However, basic but essential good practices are to be observed during a Voice UX project.

To keep things simple, we've grouped these best practices into eight major steps, from designing to finalizing a vocal experience.

Step 1: Know Your Users.

First and before doing anything, take the time to define precisely who the interface you want to design is for.

To get there, you will need to ask yourself the following questions:

  • Who is my end user?
  • What are their needs?
  • How does he accomplish them today?
  • With what words or expressions?
  • In what context or personal or professional situation?

This step is probably the most important since it will condition the voice assistant's personality, his values, the words chosen, or the turn of the questions.

Step 2: Find Your 'Tone of Voice'

The 'tone of voice,' or tonality, is the verbal and vocal identity that you will assign to the assistant.

Here too, you will need to find answers to several questions:

  • What is the objective of the assistant?
  • What will its main functions be?
  • What will be his personality traits?

All this should allow you to glimpse the character of your future assistant and his way of expressing himself: choice of familiarity or address, direct answers to the user's request or refinement by other questions, the structure of sentences confirmation/validation...

To help you, do not hesitate to define three or four qualifiers to describe your assistant and, why not, to visualize a person (real or fictitious) who could represent him.

Step 3: Define Use Cases.

The use cases, or use cases, correspond to the situations in which your interface will be used.

For this step, list these use cases and define more precisely the different functions. 
Do not try to be exhaustive and anticipate all user requests; it is simply impossible. Instead, limit yourself to the most frequently asked requests.

Step 4: Choose the voice.

The voice you choose and accompany your users for months, even years, is the most sensitive point of a voice interface project.

Fortunately, if you have followed the previous steps carefully, finding a reassuring and empathetic voice that matches your project will be much easier.

Note, however, that you have two possible options, each with advantages and disadvantages:

Human voices: more realistic and expressive; they also offer a wider choice of timbres and intonations. However, they are more complex to manage since a new registration is necessary for each change,

Synthetic voices: available in many languages, they are easy to modify and allow immediate rendering thanks to text to speech. However, they may lack expressiveness and transmit a humorous message with more difficulty, for example.

Step 5: Write the Dialogues.

At this stage, we start writing the different possible dialogues. The user's requests are anticipated and the necessary reformulations if the assistant does not find an answer or does not understand the request.

As soon as dialogue is finished, get in the habit of activating text to speech to immediately get an idea of ​​the rendering and make corrections if necessary.

Step 6: Humanize Your Interface.

The basis of a successful voice interface is the immediate understanding by the user of what the assistant is doing.

Using light or sound signals, the interface must convey that it turns on, off, is listening or performing a search.

This is the basis, but it is possible to go much further by integrating: 

  • Polite formulas,
  • Verbal confirmations,
  • Of discourse markers or logical connectors, to gain fluidity,

Error management is also critical since a lack of understanding by the assistant can have a negative and lasting impact on the user experience.

So play the transparency card when the assistant does not understand or does not know while avoiding confusing yourself with excuses not to damage trust.

Finally, be proactive by redirecting the user to alternative answers.

Step 7: Impeccable Confidentiality

Depending on the objectives of your interface, users may be required to give their contact details or their credit card number orally.

Establishing a climate of trust is therefore imperative!

In a white paper on voice assistants, the CNIL recalls the steps to follow to comply with the GDPR:

  • Identify what personal data is,
  • Clearly define the processing, its controller, and its legal basis,
  • Choose the data collected and the retention periods,
  • Inform people and guarantee their rights,
  • Protect and secure.

Step 8: Test, Correct, Improve.

We cannot repeat it enough, do not wait until it goes into production to find out what the public thinks about your product.

As soon as the voice is chosen and the first dialogues written on paper, organize test sessions to validate with your users:

  • The sound identity of your project,
  • The relevance of the scenarios,
  • The assistant provided the clarity of the answers.

While the possibilities offered by voice are colossal, this mode of interaction does not necessarily adapt to all stages of the purchasing journey.

Before diving into the deep end, take stock of the products and services you offer, their level of complexity, and their impact on your journey.



I increase mobile app engagement by developing custom voice assistants that works like “Apple Siri”. Talkit AI combines the smartest voice AI with engineering expertise to help brands build conversational voice assistants. With a customizable and scalable AI voice Assistant , we have the tools to build a highly accurate and responsive voice user interface.

Leave Comment

Please enter subject.
Please enter comments message.


Liked By