Introduction to conversation design: why, what and how

Artificial intelligence (AI) is becoming increasingly more important in our daily lives. Millions of AI-powered conversations are taking place every day, but how do these AI assistants understand what we’re talking about, and how do they know how to respond? That’s conversation design.

Conversation design is the art of mediating conversations between people and AI assistants. While the technology that powers them may differ from company to company, the science behind what makes conversations successful goes back all the way to Aristotle. Language is the most powerful tool humanity has ever created, more than 200.000 years ago.

As AI assistants, voice assistants and now AI agents become increasingly popular, it is important to understand some of the fundamentals. Sure, we can talk about advanced psychological principles that can be applied when designing for conversational interfaces, but for now, we are sticking to the basics.

It’s important to get the fundamentals right so that we can ensure that those billions of users will at least have a good and human-centric experience. A great place to start is Sample Dialogue and learning about the basic conversational structure of acknowledgment, confirmation, prompts.

In conversation design, sample dialogue is a form of role-play based on a user persona’s likely scenario. It is used to create a low fidelity conversational flow for your chatbot or voice assistant. Two people sit back to back from each other. They need to sit back to back because we don’t want any visual communication. When they can’t see each other, they only have their voice to work with. This is perfect for designing conversational interfaces.

What you’ll notice is that when we design with our voice, we naturally apply some of the basic principles that guide all conversations: turn-taking, acknowledgements, confirmations, discourse markers, and the importance of clear prompts (i.e. asking the right questions) – and that’s the power of conversation design.

If humans and AI will be living and working together, they need to learn to communicate with each other. This is where conversation designers come in. They create trust between people and AI. If you look at all technological developments in the market today, then you know conversation design is going to be an important job. And it’s not just us saying that. Here’s a clip of Bret Taylor, Chairman of OpenAI and Co-founder of Sierra, talking about the importance of the role of conversation designer.

Human centricity is at the heart of a good conversational experience. Most AI assistants are being developed by engineers, and they do not always build with the end-user in mind. These experiences often lack so-called ‘human centricity’. There is little empathy and this is causing the user to feel lost and unimportant. It increases the odds of a user dropping out of the experience. 

That's why we want to focus on a human-centric conversation design process. We want to figure out a user’s doubts, motivations, and relevant context. We want to get to the bottom of this and use it to develop an experience where users feel understood. This builds trust and allows you to take more control of the experience.

Let’s be honest, most chatbots and voice assistants kind of suck. They often don’t understand what you say and what you want to achieve. We have all been there. But some brands get it right. They know how to deliver human-centric interactions that help users get their job done when they reach out to the AI assistant.

So what do these brands have in common? If they all have access to similar technology, why are some organizations seeing more value than others? The secret: they apply essential conversation design strategies to make their chatbots or voice assistant more helpful, natural, and human-centric.

One of the best strategies in conversation design is to make sure that your AI assistant speaks in a consistent tone of voice. To do so, you need to create a persona. This makes your chatbot or voice assistant more likeable and more trustworthy. Of course, the personality must make sense for your brand and the way it relates to your users and what they are trying to achieve.

Chatbots and voice assistants can be used for repetitive processes that often can be automated. That is why customer service was the first to adopt chatbots and make them part of their regular operation.

However, conversational technology can be applied in many more ways. Think of service, marketing, sales, hr, and even online coaching. From a tech point of view, the implementation is pretty much the same for every use case. But from a conversation design point of view, different rules apply. To figure out how this works, there are two very important models: The Bill Price irritation-value Model and the BJ Fogg Behavior Model.

These are frameworks that we teach in our courses, as well as advanced writing and prompting techniques that can be applied to creating more helpful and effective AI assistants for different use cases and applications.