Natalie Atangana
Authors name: Natalie Atangana April 30, 2024
CXS Video Thumbnail (300 x 200 px) (1000 x 900 px) (1)
Conversation Design in the Age of LLMs

In the rapidly evolving world of AI, the nuances of conversation design become increasingly crucial, particularly in the context of Large Language Models (LLMs) like ChatGPT. Hans van Dam, Founder of the Conversation Design Institute, offers a wealth of insights into the intersection of technology and human communication needs. In this blog, we integrate more of van Dam's specific advice, focusing on the strategic role of LLMs, the importance of organizational structure, and practical tips for creating effective conversation designs.

Understanding the Role and Limitations of LLMs

LLMs are engineered to mimic human-like conversation, aiming for naturalness and empathy. However, as van Dam points out, LLMs are designed and trained to sound natural, empathic, and very human, but they are not designed to solve a problem, or connect the user quickly with an agent in a certain situation, or keep the user away from the agent. This is something that needs to be designed, scripted, and thought through. This distinction is vital for businesses to understand: while LLMs can enhance the way interactions feel, they do not inherently know how to navigate the strategic goals of a conversation.

The Importance of Organizational Structure and User Experience

A successful Conversational AI implementation isn’t just about technology; it’s equally about the organization behind it. You need a supportive organizational structure that aligns with the Conversational AI strategy.



“If the organization isn't structured to support ongoing learning and adaptation of bots, you're going to hit a wall"

Hans van Dam, Founder of Conversation Design Institute


This alignment is crucial for transitioning from simple bots to more advanced systems that drive real business value.

Six Practical Tips and Strategies for Effective Conversation Design

  1. Acknowledge, Confirm, Prompt

The design pattern: Ensure every interaction acknowledges what the user has said, confirms understanding, and leads the conversation forward with a prompt. This builds user confidence and keeps the conversation on track."

  1. Script Strategic Interactions

Not all customer interactions should be automated, and not all should escalate to human agents. Design your conversation paths to strategically direct users either towards self-service solutions or to live support, depending on the complexity and nature of the inquiry.

  1. Use Active Language and Simplicity

Communicate clearly and directly, using active voice to engage users and make instructions easy to follow. This approach helps prevent user frustration and confusion.

  1. Empathy by Design

Crafting responses that reflect understanding and care is crucial. It’s not just about solving a problem; it’s about how the user feels during the interaction. This can involve using language that mirrors the user’s emotions or offering reassurance when needed.

  1. Continuous Testing and Improvement

Regularly test and refine your chatbots. Use design patterns and user feedback to ensure that conversations are both relevant and effective. This ongoing process helps maintain a high standard of interaction.

  1. The One Breath Test

A message too long for a single breath probably needs cutting down. This simple test ensures clarity and conciseness, key for keeping users engaged.

The Human-Technology Skill Gap, or Why FAQ Bots Succeed and Then We Fail

The initial allure of FAQ bots lies in their simplicity and efficiency in addressing common, straightforward queries. This capability can quickly generate visible value, securing initial buy-in from stakeholders. However, as organizations ambitiously push towards deploying more sophisticated AI Agents capable of handling complex interactions, a significant disconnect emerges. This disconnect stems from the rapid advancement of technology outpacing the skill development of the humans tasked with deploying, managing, and optimizing these systems.

Bridging the Gap

For Conversational AI to truly succeed and scale, there's a pressing need for human skill development to keep pace with technological advancements. To be successful, humans need to skill up as well, to make use of the technology effectively. This isn't just about understanding how to use new tools or features; it's about developing a deeper comprehension of the strategic implications of Conversational AI and how to design interactions that leverage technology to meet complex human needs.

Strategic Implications for Organizations

This observation has profound implications for how organizations approach the development and scaling of Conversational AI. It underscores the necessity of investing in continuous learning and development for teams responsible for AI initiatives. By fostering a culture that prioritizes skill development in tandem with technological adoption, organizations can mitigate the risk of their employees lagging behind their technical capabilities.

Watch the Full Presentation from Hans van Dam at the Cognigy Experience Summit:

CXS Video Thumbnail (1)

Incorporating Hans van Dam's insights and emphasizing strategic planning, organizational support, and continuous improvement, businesses can better leverage LLMs to enhance customer interactions. By understanding both the capabilities and limits of these technologies, organizations can craft more effective and empathetic conversational experiences, ensuring that AI tools not only mimic human conversation but also strategically align with business goals and enhance the overall user experience.


"LLMs are designed and trained to sound natural, empathic, and very human, but they are not designed to solve a problem or connect the user quickly with an agent in a certain situation, or keep the user away from the agent. This is something that needs to be designed, scripted, and thought through."

Hans van Dam, Founder of Conversation Design Institute



About Hans van Dam

Hans van Dam is the founder of Conversation Design Institute, the world's leading training and certification institute for people working in Conversational AI. Hans has trained more than 5,000 professionals in Conversational AI and works with Fortune 500 companies to make their chatbots more human-centric, effective, and inclusive. He also leads CDI Foundation, a non-profit that develops design standards and provides scholarships within the industry. Hans is a frequent guest lecturer at universities worldwide. He is currently based in Amsterdam, where he lives with his family.

Other relevant resources:

How Enterprises Can Successfully Advance with Their Conversational AI Initiatives

AI Agents For Your Business

Unlock the next frontier of customer service with Large Language Models

The Future of AI: A Vision for Customer Service 2024 and Beyond