In working with major enterprises, we see a pattern in mistakes they make when designing conversations for their chatbots and voice assistants. You can't blame them... it's just how these organizations operate. Let's have a look at the challenge and see how we can avoid making these mistakes.
Focus on knowledge management
Every company has knowledge managers. They maintain the knowledge throughout your organization and these people have been working for the company for years. They are good at what they do, but these people live for the exceptions. The mindset will be to be as complete as possible in every answer they give. So if the answer is right for 80% of all people, your knowledge manager will dive in, and highlight that for 10% of the people it is answer B, and for another 5% answer C, and so on.
Knowing the exceptions is valuable, but is not a great starting point for designing conversation. We want to focus on what is going to work for most of your users. It is Pareto principle, right? The 80/20 rule. In most cases, roughly 80% of the questions come from 20% of the causes. Start covering them first and you will be able to serve the majority of your audience.
Think technology first
A technology-driven approach can also lead to bad conversation. Quite often companies start off by looking for a technology stack. Once they have invested in a conversational platform, they will start designing dialogues with that framework in mind. Programmers often seem to think that every problem asks for an (often time-consuming) technical solution. It limits their ability to design for natural conversation because they are too busy with the inner workings of your platform. It leads to a design process that is defined by technical restrictions, which limits your design process, and the creativity that you want to bring to the table.
Take the business process as starting point
The third mistake companies make, is to start designing conversation with a business process in mind. There are protocols for everything that needs to be done within an organization, and surely a business cannot go without protocols. But translating a business process 1-on-1 into a flowchart is not representative of natural conversation. Instead, we should shift our focus to the user.
A user-centric approach puts natural conversation at the heart of design. And how do you do that? We create sample dialogues together. It is a kind of improvisation acting, really. It is role-playing. We put two people in a room, establish what a user wants to know and what a bot can and can’t do, and let these people have the conversation for us. Record and document your conversations and you will have a solid basis for designing natural conversation.
Of course, this is just the start. By applying different copywriting techniques, that take both psychology and technology into account, we can create and shape meaningful dialogues that are helpful, natural and persuasive. In our Conversational Academy, you can learn everything about sample dialogues, WOZ-testing, Happy Conversation Design, and much much more.