Agent Experience: Building an Open Web for the AI Era

Market Thesis: AI acts as the primary disruptor for the web, driving two predictable shifts: accelerating the volume of code/content creation and enabling "UI 2.0" (fundamentally new, creative experiences impossible previously).

• Cost Convergence: The marginal cost of creation (code, assets, video, 3D) is converging toward zero, removing previous economic barriers to complex media like WebGL and 3D rendering.
• Predictability: No industry participant predicts slower development; the consensus is that AI will universally increase throughput.

Strategic Pivot to AX (Agent Experience): Netlify is shifting its core philosophy from DX (Developer Experience) to AX (Agent Experience), defined as the holistic interaction between autonomous agents and the web platform.

• Definition: AX extends Don Norman's UX concept to agents (AI systems) that write code, generate assets, and collaborate with humans.
• Goal: Prevent the web from becoming a "content repository" for walled-garden AI platforms (like TikTok) by making the open web the native, preferred environment for AI agents.
• Interoperability: The web must support agents interacting with pages in various modes (e.g., data extraction, screen-reader style, or full interactive rendering) without requiring closed ecosystems like Flash.

Current Usage Metrics & Trends:

• Scale: Netlify now processes approximately 10,000 sites created daily via AI code generators (e.g., Bolt.new).
• Adoption Spectrum: Usage spans non-technical users (democratizing app building) to professional developers using AI to replace boilerplate and frameworks (e.g., using Cursor, Bolt, or Windsurf to skip npx setup).
• Emerging Developer Base: AI tools are creating "new programmers" who learn by prompting and iterating on generated code, potentially adding 100 million new web developers over the next decade.

Technical Primitives & Architecture Evolution:

• LLM-Friendly Code: Developers are rewriting codebases (e.g., switching from React Hooks to Relayed single-file components) to expose more semantics for LLMs to reason about, improving "one-shot" generation accuracy.
• New Infrastructure Primitives: Traditional infrastructure is evolving to suit ephemeral, agent-driven workloads.
  • Serverless Databases: Concepts previously considered "unnecessary" (scaling to zero) are now viable due to the rise of temporary, AI-generated applications.
  • Ephemeral Apps: The ability to "fire up" an app to solve a specific problem is driving demand for infrastructure that supports instant creation and termination.
• Native Primitives: The web needs new native primitives to replace current "headless browser" and screenshot-based agent interactions, aiming for direct, structured agent-to-platform communication.

Future Outlook & Creative Potential:

• Creative Democratization: Lowered costs enable individual creators (e.g., TikTok creators, solo devs) to produce high-fidelity content (sci-fi epics, 3D tools) that previously required Hollywood budgets.
• Platform Choice: The objective is to ensure these new creative layers and agent interactions remain on the open web rather than migrating to closed AI interfaces.
• Skill Shift: Future development roles will bifurcate: deep-stack expertise for building core systems vs. "agent-prompting" and vision for the next 100 million developers.

Discipline & Standards:

• Netlify launched the "Resource Agent Experience Study" to foster cross-industry collaboration on defining standards for agent capabilities.
• AX is proposed as a formal discipline (similar to UX/DX) requiring intentional design of authentication flows, documentation, and primitives specifically for autonomous agents.