How To Design Amazing Conversational Interfaces & Chatbots

A design thinking approach is crucial in helping designers figure out what users want from their products. But what if you’re designing a chatbot or a voice assistant in which conversation is the interface?

There are two types of conversational interfaces – chatbots, which you type to, and voice assistants, which you talk to. To illustrate, you type to Mastercard’s chatbot to locate the nearest Subway outlet and order a sandwich online. And for Siri, a voice assistant, you can summon verbally on your iPhone and ask about today’s weather.

In this article, I will share some tips for designing amazing conversational interfaces and chatbots. So let’s get started.

1. Add a Touch of Empathy

Human interactions are taken for granted, but if you pay attention to them, you realize how intricate they really are. The way knowledge is transferred, the different ways people talk and express social awkwardness, all communicate their intent and make social exchanges more comfortable.

To make AI applications offer the same experience, add a touch of empathy.

Empathy is the heart of great design. In conversations, we express empathy through language.

And since interactions in AI chatbots are purely conversational, it makes sense to add a touch of empathy in their design by considering the following:

Understanding Context:

Imagine saying “I am depressed” and getting a reply “I hate to hear that” like Cortana does when asked the same? It’s an insensitive answer and won’t make you feel very good. And worst of all, it doesn’t really help your situation.

To avoid such a scenario, offer to help by providing with the contact details of nearby medical experts. It’s not hard to target keywords such as ‘depressed’ or ‘chest pain’.

Keep your script interactive and conversational:

A conversation is like a dialogue that make people comfortable enough to participate. Statements read in prose, on the other hand are regarded with disinterest.

To make the copy of your conversational UI engaging, make it talk like you would naturally do with someone else. Keep this in mind when writing the script for your conversational interface.

How? Here’s a start. Read what you wrote out loud. Would you ever say this to a friend? Does it sound casual? If your answer is yes, then you’re on the right track.

Also, a good thing to do at the beginning is to introduce your bot to the user. This will set some expectations, and the user will respect its capabilities and limitations..

2. Make it visual

Sometimes, words aren’t enough. To get your point across and put more emphasis on what you have to say, you have to resort to visual cues. Studies have shown that people engage more with visual content. One of the reasons why we use graphs and charts to display reports in business meetings.

In the case of chatbots, visuals can be helpful to elicit certain responses. Here is how you can incorporate them in your conversations:

To trigger emotions:

We use images to convey feelings, to deal with situations that cannot be dealt with words alone. For example, you may use a smiley emoji in a welcome message as it evokes a warm response and a subtle, yet positive difference to a conversation.

To offer help:

In some cases, visual cues are intuitive enough to offer help. The symbol on a restroom door clearly indicates whether it is intended for men or women. With this in mind, train your conversational interface to use images when the need arises. For example, if a user asks for directions to a certain location, your application can present him with a map.

To emphasize on the importance of visuals in conversational interfaces, consider the Whole Foods Market chatbot. After you select an emoji, such as a banana or an an eggplant, it sends you a recipe that has this item from it’s database.

3. Prepare for all possible use cases

What’s common between good chatbots is how well the use cases their script covers. Sit down with your team, and use mind mapping techniques to jot down all possible scenarios that a user may respond with and your plan to deal with each one.

Imagine a user asking a question on which a bot answers with ‘I can’t understand what you’re talking about’ or perhaps ‘Can you please explain it further’ multiple times during the conversation. It will be a major upset and you’re most likely to lose a valuable user.

A good script should help you create different scenarios of the conversation. Hence, make sure that no dead ends are left in your chatbot’s script.

4. Aim for Convenience

Conversational interfaces like Google Assistant and Messenger bots are designed to offer convenience. A business that uses a chatbot for its customers offers them the convenience of ordering items without having to step foot in a store. You can do the same. Here is how:

Be Clear and Specific:

Keep in mind, while conversational interfaces do mimic human properties, they have their limitations. You can’t interrupt them to gain clarification like you would with a customer support executive.

Hence, to minimize any possible confusion, your messages need to be concrete and clear. A response shouldn’t contain two messages that are totally unrelated to a user’s intent. For example, if the intent is to make a restaurant reservation, the call to action should only have buttons that acknowledge it such as Table for two? Table for four? etc.

Keep User Input to a Minimum:

The less users have to interact with your interface, the more convenient the experience. Keep interactions to a minimum by providing users with only what they need to move ahead. Don’t offer multiple options to a question when a simple “Yes” or “No” would suffice.

Wrapping Up

Conversational interfaces, both chatbots and voice assistants, have brought with them a new type of interface. If developers hope to create conversational interfaces that offer AI experiences which are truly intuitive, they must think out of the box by:

• Empathizing with users and keeping the language natural
• Using visuals to trigger emotions and offer help when needed
• Preparing for edge cases
• Aiming for specificity and keeping user input to a minimum to keep the experience convenient

Next: Enjoy reading about Artificial Intelligence? Find out if we are headed into a future of chatbots. Considering the innovations made in the field, it might not be long before conversational interfaces become a large part of software development.