Why Conversation is Important in Designing Voice UI?

Rittam Debnath
NYC Design
Published in
7 min readApr 9, 2020
Image Credit: Gleb Kuznetsov

The sense of touch and pressure has been the common form of interaction in terms of Human-computer interaction for the past few decades but one thing which we forget is that there is another more natural and intuitive way of interaction which is our voice.

Voice UIs have gone standard and are changing lives and expanding availability for some consumers. At a similar time, the present voice user encounters (most popular as Cortana, Alexa, and Google Home) remain established in an extremely basic, order and-control procedure. We can just call the present encounters “conversational” in the broadest feeling of the word — as verbally expressed words exchanged.

There is a lot of skepticism around the field of Conversational and Voice Design. At least this was the case when I first started working on a voice project, we were a team of myself — and a few other UX researchers. Very soon we learned that conversational design is like creating chatbots and voice UI design is an infusion of both conversational design and to create an interaction flow between a voice assistant and a human.

Comparing to other projects that I have been a part of is more or less the same, you recruit participants, do a contextual inquiry, gain some understanding of what their pain points and needs are and then you move on to create a persona and finally skinning the product. Now coming to the Voice project it is not the UI you are trying to create its the conversation; here you need to have a good writing skill.

In voice, writing is design.

You must have a few good writers in your team who can use their creative thinking and increase the quality of voice applications.

What the future of Voice Interaction looks like

Now Voice Interaction could be divided into two parts:

  1. Order / Command
  2. Control / Response

One thing which we lack in Interaction design is the Response phase. Making a Natural Language Voice Assistance(NLVA) in a foreseeable future is not a very tough task because people around the world are helping to build it but the question is, will it be able to match JARVIS? Whenever Tony Stark asks a question to JARVIS its not just the question that the NLVA analyses but the subsequent relation to the question and its past occurrences. It remembers what a human is saying and empathize and understands to give a relevant and correct answer.

Human Interaction and Voice

In the TED talk by Don Norman, he talks about how emotion is all about acting. Condition is about understanding the world, emotion is about interpreting it — saying good, bad, safe, dangerous, and getting us ready to act. he also talks about behavioral design which is all about feeling in control, which includes usability, understanding — but also the feel and heft.

This is the reason why Voice assistance lacks the human touch. NLVA shouldn’t react to the question that we ask instead it should understand and then prepare a response so that it becomes a conversation.

Have you at any point truly attempted to separate the steps? So consider sitting with your friend at a bistro and simply having a casual talk. How would you realize when to go ahead? How would you realize when to move on to the next topics? Furthermore, how would you even know what subjects to talk about? It’s just plain obvious, the greater part of us don’t generally consider it since it’s practically natural. What’s more, when we find a good pace, we get familiar with what matters to them, and afterward, we realize what points we can explore. Be that as it may, with regards to showing AI frameworks how to communicate with individuals, we need to show them bit by bit what to do. What’s more, at present, it feels inconvenient.

If you’ve at any point attempted to converse with Alexa, Siri or Google Assistant, you can tell that it or they can, in any case, sound cold. Also, have you at any point get irritated when they didn’t comprehend what you were stating and you needed to rethink what you needed multiple times just to play a song? Okay, to the credit of the designers, realistic conversation is extremely hard. What’s more, there’s an entire part of humanism, called conversational analysis, that attempts to make blueprints for various kinds of discussion. Types like customer care or advising, instructing, and others.

Conversation Analysis

If you want to know more about conversation analysis there is an interesting concept introduced by philosopher H. Paul Grice in his 1975 article “Logic and Conversation” in which he suggested that meaningful exchange is defined by cooperation.

“Each participant recognizes in them, to some extent, a common purpose or set of purposes, or at least a mutually accepted direction.

Best practice to build a conversation in Voice UI

Conversation design is a pretty new approach which came into popularity due to worldwide advancement in voice assistance technology. But no matter there is always a point which might turn into an awkward conversation. Accordingly, to be a great conversation designer you need to be a great conversationalist. Here are ways to design a VUI that talks more naturally.

Purpose gives direction: Have you tried to have a selfless conversation and felt completely lost?. To make sure this doesn’t happen you have to know what is the purpose, which in turn provides direction that leads to more engaging conversation.

People like being heard: Every healthy relationship is built upon trust. If we make VUI’s listen to users carefully and acknowledge that they have been heard by confirmation it gives the user an emotional satisfaction that someone is listening and builds the connection.

Both parties should be cooperating to push the conversation: Sometimes users say the answer prior to the predictable question that the bot is gonna ask.

Check for Conversation Errors: As a human, we make mistakes while talking but when we talk to another person we also throwout facial cues that helps repair the conversation from being awkward. But in the case of VUIs, it doesn’t receive any visual cues so it is the job of designers to create conversational repairs. For instance.

Did you mean,______________?

And also Rapid Re-prompts, asking a short immediate question for instance:

Sorry, what was that?

Words carry emotions, So Design VUIs that convey a personality that seems more human.

How do you design for Voice?

First off, different platforms have different names to their AI Assistance, For instance, Amazon calls them ‘Skills’ whereas Google chooses to call them ‘Actions’. Every platform has its own way of programming these skills and also various in-built features. Alexa is more blessed in this case cause Amazon allows developers to integrate third-party skills easily when compared to Google.

Building on Amazon Alexa

You can either use the Alexa kit or you can use VoiceFlow. A little bit about VoiceFlow, it is a SAAS platform which allows user to create VUI’s without coding also allows you to integrate the conversation to Alexa and google home. This makes it much easier to create a sample VUI within few minutes. Users can just drag and drop the components on to the board and connect this component to create a conversation flow.

Recently I was a part of a team where we aimed to explore how a technology-based solution can help people search for dining buddies.

During this process, we thought what if we can do the same thing by asking Alexa or Google to search for your dining buddies. How would the experience change? Here is an example.

Voice UI for Dining Buddies

Putting it all together

Conversations and to a bigger degree, language, is intricate and loaded up with subtleties. Be that as it may, this isn’t to imply that we can’t improve our capacity to take part in better discussions.

The objective isn’t to reproduce human collaboration. The objective is to inspire a connection with a discussion that feels less computerized and helps users with what’s proposed.

The best discussions cause individuals to feel great, adjusting to the meek changes immediately and dialogue flow normally. Soon individuals will consider voice a characteristic method for interfacing with products and services. Simply envision that in ten years, you’ll stroll into the house and control a wide range of machines using voice. By concentrating on building trust, designers will make frameworks that will be utilized to respond to the most significant questions, for example, “What would my health condition be after 5 years based on my current food consumption?”.

Check out my portfolio.

P.S. If you have any feedback hit me on my website. If you like this, make sure to share and recommend (by clap👏👏👏 ).

--

--

Rittam Debnath
NYC Design

Design Conversationalist. I write about Design in Human Interaction. Working on my next article. Love to connect with you. Connect with me: rittamdebnath.in