Conversational Technologies
4 min read

The Art of Effective AI Assistant Conversation Design

Sharona Reouveni
Sharona Reouveni Conversation Designer
The Art of Effective AI Assistant Conversation Design

Nowadays, more and more companies want to use technology to automate communication with their customers on different channels. Deciding which conversations to automate and which to leave to human agents is critical. 

Technology can’t replace the human touch in high-risk, sensitive, or emergency situations. However, repetitive conversations concerning appointment management, prescription refilling, or payment and authorization can be automated to lighten the load on human agents and improve the patient experience. The right solution can make these conversations feel more natural and enjoyable.

Hyro’s ability to turn complex data into simple dialogue allows our customers to automate conversations using our web and voice assistants. But automating interactions is only the beginning. We add our knowledge and understanding of the healthcare industry, empowering our customers with best practices collected and perfected over time to enrich and optimize dialogues. One of the ways Hyro achieves that is through conversation design.

What Is Conversation Design?

Conversation design is a new and constantly evolving discipline focused on creating human-centric and value-driven AI assistants. As Hyro’s conversation designer, I aim to create conversational flows that are helpful and pleasant. In other words, I translate business and product requirements into natural conversations that bring value to our users and customers. 

To be effective, our assistants must be clear, empathetic, and able to adapt to different situations and channels. To this end, I use recipient design and user research methodologies and curate standards and best practices. I also develop infrastructure and reusable design patterns that allow Hyro to maintain and scale this process. Improving conversational user experiences (CUX) requires close collaboration with our engineering and product teams to align our goals and technology. 

From Blueprint to Reality: The Process of Designing Conversational Flows

I start designing a new flow by collecting requirements. Then, I jot down a happy flow—imagining an interaction between Hyro’s virtual assistant and a user, where everything goes well, and the user easily gets what they need. I turn that into a conversation flow, adding edge cases that come up as I progress. We then test the flow with the product and engineering teams to ensure it meets the requirements. 

My work continues after the new skill is launched. We follow up with an evaluation process, collecting feedback and data to improve the flow and add any missing elements. In addition, we constantly monitor our conversations, using different conversation analysis processes and mechanisms to maintain and improve our user and customer satisfaction.

A Systematic Approach to Conversation Design

Conversation design harnesses the powers of psychology, linguistics, and large language models (LLMs) to analyze, streamline, and improve conversations. We use different tools and methodologies to minimize cognitive load and structure conversations in the right order, format, and timing. Ethics is also a significant consideration, especially when dealing with healthcare. We must protect the patient’s privacy, restrict personal and medical information use, and adhere to industry regulations.

While giving our customers an easy-to-deploy out-of-the-box solution, we also allow them to customize our AI assistants to their tone of voice, style guides, working procedures, and organizational guidelines. That is another consideration always on my mind when designing conversation flows. I consider codeless customization options, empowering our customers to adjust dialogue to their and their users’ persona. 

Conversational AI for the Real World

People don’t like virtual assistants—a bias stemming from years of dealing with inflexible, robotic, and unhelpful chatbots. Hyro is on a mission to change that perception by putting our users and their needs at the center. Our conversational AI assistants accommodate themselves to the users, dynamically handling curveball questions and transitions within skills.

The healthcare industry caters to a widely diverse audience in sensitive circumstances. Our users may feel overwhelmed, worried, or in a rush. They include both native and non-native English speakers, both patients and caregivers. Our assistants need to adjust to different needs, intents, and situations. 

Impactful Conversations that Make a Difference

Being helpful is about reading the room, which we aim to do with our linguistics and LLM capabilities—understanding what the user needs and helping them get there as smoothly as possible. We’re not trying to imitate a human agent but focus on the human experience of our users. When all’s said and done, if our assistants helped improve someone’s day by relieving some stress or dramatically shortening waiting times, freeing them to care for themselves or a loved one, that’s our job well done. 

If you’re curious to learn more about our conversational AI platform and how we put human-centric conversations into action across some of the largest health systems in the US, sign up to get more articles like this below or speak with one of our experts.

Conversational AI insights,
directly to your inbox.
About the author
Sharona Reouveni
Sharona Reouveni Conversation Designer

Sharona is Hyro's Conversation Designer and an accomplished UX writer. A master in untangling complex systems and optimizing user experiences, Sharona first started crafting dialogue as a certified puppeteer, which can totally be considered conversation design under very particular circumstances.

Continue Reading
See all blog posts

Continue Reading

See all blog posts