Why is Conversation Design Important?
We’ve all been there. We’ve typed in the URL, hit enter and waited for the site to load. A text box pops up, usually on the bottom right-hand side of the website. The bot introduces itself and offers to resolve any queries we might have. Now, at this stage, usually one of the two things happens. Either we hastily close the pop-up, owing to the previous dreadful experiences we’ve had with such conversational interfaces. Or we proceed to give it a try, knowing fully well that the chances of resolution of our query are bleak, and we’ll eventually be spending time scouting the information ourselves anyways.
Why such dread?
Despite the recent advancements in conversational technology, most of the conversational interfaces we encounter daily fail to hit the mark. They promise a lot but fail to satisfy even the basic demands. They misunderstand our queries, bombard us with repetitive answers and take us around in circles. Consequently, the interactions turn out to be taxing and unfruitful. And since people tend to remember the bad more than the good, poorly-designed conversational interfaces always stick out and give the whole industry a bad name. That is why the vast majority of us tend to be sceptical rather than appreciative when interacting with any conversational interface.
One of the reasons for a poor conversational experience is that organizations tend to inadvertently use a technocentric approach as opposed to a user-centric approach. The choice of platforms and APIs seems to overshadow careful user research. They are more concerned with how a particular conversational AI solution will integrate into their website or app rather than how will it affect the overarching user experience. Instead of carefully evaluating the use case and focusing on value addition, organizations spend time integrating as many features as possible to make the bot more “smart” and “capable”.
Another reason is the ubiquitous nature of conversational AI platforms. For a long time, implementing conversational interfaces into existing solutions was a mammoth task and required a fair bit of technical expertise. Organizations spend a lot of time on the technical side of things. But now, with the recent advancements in the conversational AI field, the technical complexities have been abstracted away. Conversational AI platforms now offer out-of-the-box omnichannel capabilities, multilingual options and self-learning capabilities. This makes it easy for organizations to incorporate conversational interfaces within their offerings. So naturally, the number of conversational interfaces has risen significantly over the past years. But again, the nonchalant outlook towards design has resulted in more and more poorly designed conversational interfaces.
What is a well-designed conversation anyways?
A well-designed conversation just seems to work. Inseparable from magic, it checks all the boxes of a successful machine-human interaction. The interaction feels empathic and is centred around the user. It is concise, yet effective. The cognitive load is minimal. The responses are natural and contextual. Bloated words and technical jargon are skipped in favour of terminology catered towards that user’s vocabulary level. The responses are reminiscent of the brand. The bot comprehends even mistyped or mispronounced words. There are no discernible breakpoints in the entire conversation. There are no endless loops either. If the user’s queries aren’t answered, the conversation gracefully transfers to a real person. Or, he’s pointed towards the right direction. It is all-natural and just seems to work.
Conversational interfaces can tremendously reduce the cognitive load for any given task. And that itself acts as a litmus test. If the conversational interface requires more mental effort than a traditional user interface, there is a chance that it is a poorly designed one. And since the field of Conversation Design is new, it is hard to pinpoint exactly where one goes wrong. Organizations must realize that efficient automation does not automatically translate to a successful user experience. Conversational AI might reduce the load on their customer support teams, but at what cost?
What are the elements of a good conversation design?
There is no single template for a well-designed conversation. However, certain crucial elements are instantly recognizable in well-crafted conversational interfaces. Organizations need to be mindful of them during the design process. Some of these elements are —
- User Goal — A user interacts with a brand with a goal in mind. It can be query resolution, customer support, or general enquiries. Hence, the design should be mindful of the end goal. The bot responses should be clear, concise and fact-based. With each turn of the conversation, the user inches one step closer to his goal.
- Persona Design — A conversational interface should be reminiscent of the brand and should feel like an extension of the existing brand identity. This helps establish trust and mimic existing user relationships throughout the interaction. Having a bot persona helps maintain consistency throughout the interaction. The responses can be formal and factual for a business-facing company, while fun and light-hearted for a more consumer-friendly company.
- Clear, concise and multiturn — Brevity is the soul of wit. A well-designed conversation ditches fluff in favour of clear, concise and factual information. It also facilitates turn-taking. This ensures that the conversation remains goal-oriented and natural.
- Empathic — Empathy is what separates a conversational AI from chatbots of the past. If the interaction is empathic, the user feels heard. The goal of the conversation should be less about filling slots and more about understanding the user’s situation and phrasing responses accordingly. The bot should not be responding with witty one-liners when the user is desperately searching for time-sensitive info. The phrases should be designed for different user scenarios accordingly.
- Humane with robust error tolerance — Because there is no such thing as errors in conversations, it is important to build a robust error tolerance mechanism. “Sorry, I didn’t understand that” phrases should be retired in favour of more user-centric responses. If something is out of scope for the bot, it simply escalates the query to a customer service agent or points the user in the right direction. The bot’s goal should be to help the user and not showcase its AI prowess.
Conclusion
The above list is by no means an exhaustive one. It simply tries to shed light on the crucial role a good conversation design plays in the bot-building process. This newly minted field lies at the crossroads of language, linguistics, psychology, copywriting and UX research. This is where a Conversation Designer comes into the picture. He understands the implications of the abovementioned disciplines in designing successful conversational interactions. Just as UI designers design graphical interfaces, conversation designers craft conversational interfaces. And since organizations task specialized people for designing user interfaces, shouldn’t the same rule apply to conversational interfaces too?
And if the past is any indicator, we’ve seen what happens when organizations try to skim through the design part. It sidelines the user in favour of the bot. There are contextually irrelevant responses and a bunch of “Sorry I didn’t get that. I’m still learning.” phrases sprinkled throughout the conversation. This kind of semi-nuanced, last-moment take on conversation design results in an irksome experience that consistently fails to hit the mark. That is why the role of conversation designers is becoming increasingly important in this ever-evolving field.
With an ever-increasing focus on the importance of a good conversation design, one can find numerous resources on the whys and whats but seldom the hows of designing great conversations. And as a conversation design studio, our core strength lies in crafting great conversational interfaces. So, if you need help building one, or even fixing one, ping us at hello@voiceano.com 👋😄
