Designing for Voice-First and Desired Outcomes

Uncategorized Dec 09, 2020

Hans van Dam has always possessed a creative side and a will to innovate. Once an aspiring novelist, he later copywrote for big tech companies. After a startup company failed, Hans moved on to work in customer service before shifting his focus to chatbots. Through his later work, Hans learned voice design and began to dream even bigger. He eventually started a company with the goal of training and certifying designers from around the world. Today, Hans is the co-founder and CEO of the Conversation Design Institute

Certification and training

Despite some early skepticism, the Conversation Design Institute has thrived as the need for training has grown. Hans sees a need for more structure across job titles, skillsets, and team interactions in his industry, and one of his goals is to improve alignment in these areas.

The institute currently recognizes and provides certification for three different roles: AI trainers, conversational designers, and conversational copywriters. AI trainers take on a technical role that requires an understanding of light data science. Conversational designers often have a background in UX design, and the work is research-intensive. Conversational copywriters must be strong writers who are able to turn words into effective dialogue. Although relevant experience is useful, Hans’ students “come from all walks of life.” 

What does it mean to write for listenability?

Hans views writing for listenability as the best approach to take when writing conversations. “Everything we design—it doesn’t matter if it’s a chatbot or a voice assistant—we design for voice-first.” Hans’ reasoning is that if you can design for voice, “the most complicated interface in the world,” then you can write for any application. 

Structuring an effective message in voice

There are several fundamental techniques used when structuring a message in voice. The first is to acknowledge a user’s request. Humans often give non-verbal cues to signal acknowledgement (such as a head nod), but AI might say something like, “got it.” The next step is to confirm the request through repetition. AI should end with a prompt to indicate that it’s the user’s turn to perform an action. Hans says that the best prompts are simple. They can even be a single word. Good prompts are more likely to result in desired outcomes. 

Teams should also map out conversations to take both user and bot needs into account. To be helpful to the user, a bot requires a list of rules and tasks. User needs are comprised of both informational and emotional components. When first building out exchanges, Hans’ team members even sit back-to-back and will improvise sample dialogue. Early interactions are messy—“but we’ll learn a lot.”

Balancing AI and user needs

Despite the importance of building emotionally rich, or empathetic exchanges, many of Hans’ clients start out with a robust team of engineers and only one writer. This can result in an end product that is technically sound, but unable to keep users engaged and motivated. To address this imbalance, Hans’ company builds equitable departments “where designers have a fair seat at the table with the engineers.” 

“The psychology of language and the technology always work together,” Hans elaborates.
A lot of companies fail on their projects because they don’t understand that balance.” 

Using psychology for desired outcomes

The psychological concepts of scarcity, social proof, and anticipatory enthusiasm, to name a few, can all be leveraged in design. Copywriters have a toolkit of around 50 different techniques that can be used in specific situations.

Hans gives the example of AI helping a user determine if they want to purchase a specific pair of shoes. A voice assistant will state the price, but may also mention that the shoes are popular with a lot of park-goers in the city. This use of “social proof” may influence the user to make the purchase.  A user’s ability and motivation to perform an action must also be considered. If these factors can be improved with AI assistance, then the prompt is more likely to elicit a desired outcome.

Looking toward the future

Hans believes that AI will become increasingly multi-modal, using a combination of voice, screens, and even holograms. Hans sees conversational design as a significant part of that future. “Just like you cannot imagine a company without engineers today, you will not be able to imagine a company without conversation designers five years from now.”

For those starting in his field, Hans’ advice is simple: “Don’t wing it.” He hopes newcomers seek out proper training and focus on building a portfolio for greater exposure. It ultimately takes “a combination of creativity and craftmanship” to succeed.


Sign up for the Digital Assistant Academy newsletter

Subscribe to get our latest updates and offers by email.