The UX of AI: Making AI-Powered Apps Your Users Don't Hate - Kathryn Grayson Nanz, Progress Software

2026-07-18 · Source: AI Engineer · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, extended

Summary

Kathryn Grayson Nanz of Progress Software highlights the significant user experience (UX) challenges inherent in AI-powered applications, stemming from AI's non-deterministic nature and the wide knowledge gap between developers and average users. Drawing parallels to the Macintosh UI's evolution, Nanz argues that current AI interfaces are akin to "System 3," requiring careful design to meet users where they are. The presentation outlines five critical pillars for effective AI UX: Trust, Clarity, Control, Transparency, and Meaningful Benefit. Specific strategies include citing references and human-in-the-loop for trust, streaming text and chain-of-thought for clarity, emergency stop buttons and version history for control, clear permissions and cost estimates for transparency, and guiding users with examples and integrations for meaningful benefit. Nanz emphasizes that AI cannot yet design novel AI UX patterns, making human-centric design crucial for user adoption and satisfaction.

Key takeaway

For AI engineers and product designers building AI-powered applications, you must prioritize user experience to ensure adoption and satisfaction. Focus on implementing the five pillars—Trust, Clarity, Control, Transparency, and Meaningful Benefit—to bridge the developer-user knowledge gap. Your designs should empower users, not overwhelm them, by providing clear feedback, control mechanisms, and actionable next steps, ultimately making AI a valuable tool rather than a source of frustration.

Key insights

Effective AI UX bridges the knowledge gap between developers and users by prioritizing trust, clarity, control, transparency, and meaningful benefit.

Principles

• AI UX must evolve with user literacy over time.
• Honesty about AI limitations builds user trust.
• AI cannot yet design novel, high-quality AI interfaces.

Method

Design AI features by starting with user problems, then applying the five pillars: Trust, Clarity, Control, Transparency, and Meaningful Benefit, adapting existing UI patterns gradually.

In practice

• Cite sources for AI-generated content.
• Implement emergency stop buttons for AI actions.
• Provide examples and templates for prompts.

Topics

• AI User Experience
• Human-AI Interaction
• Trustworthy AI
• UX Design Principles
• AI-Powered Applications
• User Control

Best for: Software Engineer, AI Engineer, Product Designer

Related on AIssential

• A Framework of User Experience Principles for Human-AI Agent Interaction in the Workplace — Establishing UX principles for human-AI interaction in the workplace is crucial for trust and adoption.
• AI for Everyone, Not Just for Those Who Already Know How to Ask — AI systems need inherent transparency and trustworthiness from inception to prevent manipulation and ensure equitable access for all users.
• Measuring UX When AI Is Unpredictable: New Metrics — AI UX requires new metrics beyond task completion to capture unpredictability, trust, and actual user outcomes.
• AI Can Generate Interfaces. But Can It Create Experiences? — AI generates interfaces efficiently, but human designers remain crucial for creating empathetic, strategic user experiences.
• 4 tips for building better AI agents that your business can trust — Building trustworthy AI agents requires rigorous measurement, cross-functional collaboration, and leveraging existing proven capabilities.
• The Problem With AI Everyone Is Okay With — Distinguishing AI accommodation from flattery is crucial for designing trustworthy personalized systems that preserve truth and user autonomy.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.