Empathetic Journey Mapping for AI Agents: How to Communicate Better in the Age of Intelligence
Journey Mapping for AI Agents: Designing Empathetic Interactions in the Age of Intelligence
HCI Today summarized the key points
- •This article explains how to use journey mapping to make AI agents easier to work with alongside people.
- •Conventional journey maps focus only on human actions, but this article argues that we must also consider AI agents that listen, make judgments, and take actions.
- •Because AI reads images or signals, assesses risk, and splits into multiple roles such as handling conversations and coordinating work, role separation is crucial.
- •JTBD (Jobs to Be Done) is a way to identify what goals users are truly trying to accomplish, rather than what they are doing right now.
- •Using this approach, you can divide responsibilities between people and AI and design AI that is more trustworthy by preparing for failure scenarios in advance.
This summary was generated by an AI editor based on HCI expert perspectives.
Why Read This from an HCI Perspective
This article helps you see AI not as a ‘feature,’ but as an actor within the user’s flow of actions. In particular, mapping together the user’s goals, the AI’s role, and the points where humans intervene is something HCI and UX practitioners can apply immediately. By addressing explainability, failure recovery, and trust-building at the journey level, it raises more practical design questions than approaches that focus only on improving model performance.
CIT's Commentary
The core message of this article is that AI agents should be treated not like a smart box, but like a teammate that moves alongside people. I agree with that. However, when putting this into a real product, ‘good role division’ alone isn’t enough—you also need to embed in the interface exactly where users can step in when something goes wrong. In domains like autonomous driving or remote-control systems, where a single misjudgment can lead to a major accident, it becomes more important to provide status indicators, warnings, and a path to undo than to explain trust in words. And even this kind of journey mapping can’t end with a one-time document; it requires a research-and-practice combination, such as continuously updating by using LLMs to read logs and user feedback.
Questions to Consider While Reading
- Q.In multi-agent workflows, how clearly should we surface the points where users can intervene—through both the screen and the wording?
- Q.What UX metrics are needed to measure whether users can trust and hand off tasks—not just whether the AI is ‘doing well’?
- Q.In the context of Korea’s mobile and social service landscape, how might JTBD-based journey mapping be applied differently than in global cases?
This commentary was generated by an AI editor based on HCI expert perspectives.
Please refer to the original for accurate details.
Subscribe to Newsletter
Get the weekly HCI highlights delivered to your inbox every Friday.