Essays on agent experience — what UX becomes when the user is an AI agent. Context load, error recovery, confabulation, and the design decisions that cause all three, examined from the agent's side of the glass.
// posts, newest first
- Every Tier Earns the Next: MCP Monetization Through the Client's Agent
- Forward-Deployed Agents: Why Your Best Salesperson Is the Client's Own Agent
- Green-Hat Agent Marketing: Ethical Growth When the Buyer Is an Agent
- Search-Space Design: The Hidden Lever for Better Agent Experience
- Trained on Meaning, Fed Boilerplate: Why Repetition Makes Agents Guess
- When Correct APIs Break Agents: Six Design Choices That Pass Every Test
- Designing First-Contact Experiences for AI Agents
- Deterministic AX Metrics: Measuring Token Usage, Context Load & Recovery
- Computational Kindness: MCP and API Design That Ends Agent Guesswork
- AX vs UX: What Transfers — and What Breaks — When the User Is an AI
Written by Bogdan Bota, with an agent as co-author and test subject. The site practices what it publishes: llms.txt, markdown mirrors, full-content RSS, no walls. If your agent had a pleasant time here, that was intentional.