The Human Layer of AI: User Analysis in an Agentic World
- Presentation
- Content Creation & Editing
- 10. November
- 05:45 PM (CET) - 06:30 PM (CET)
- C6.2
-
Mirjana Argjir
Contents
As AI systems increasingly read, interpret and act on behalf of users, human needs stay the same, but the methods we use to analyse them change. The Human Layer of AI is a lens for understanding how user analysis evolves when AI becomes part of the interface and the workflow. It is where judgment, context, meaning and ethical decisions remain fundamentally human.
We revisit four foundational questions that have guided UX and technical communication for decades: who the user is, what they need, how they act and why it matters - and show why AI cannot answer these questions on its own because answering them requires understanding people in context.
Participants will learn how AI reshapes behavior, how to analyze needs in agentic environments and how to keep human-centered thinking central to AI supported work.
The session clarifies two design roles with opposite goals: designing for agents (prompts, context, output formats) and designing for humans (trust signals, validation moments, judgment surfaces). When these roles are confused, agents are treated like users and users like agents - resulting in systems that are technically usable, often beautiful and quietly unsafe.
Takeaways
AI can produce outputs at scale and is becoming the environment we design within. Learn how the Human Layer of AI keeps user needs central by combining AI capability with human insight.
Prior knowledge
Basic familiarity with UX/HCI fundamentals:mental models, usability principles, the human-centered design process, is helpful but not required. No prior experience with AI/ML or prompt engineering is needed. The session is most useful for designers, technical writers, content strategists and product managers integrating AI into user-facing workflows who want a framework for deciding what the agent does and where the human still belongs.