Adobe Agentic Sites (Audience of One)
www.latent.space- Category
- AI Tools
- Pricing
- Paid
- Type
- ARTICLE
- Builder
- @swyx
- Added
- Jul 21, 2026
About
An experimental Adobe concept where websites assemble themselves in real time around each visitor's intent. Instead of static pages, an LLM retrieves from a company's existing content and composes a personalized page on the fly for individual users and even AI agents.
What it can do
Compose a personalized web page in real time based on visitor intent
Visitor intent signals and browsing context → Dynamically assembled personalized web page
Retrieve relevant content from a company's existing content repository
User query or intent and company content library → Relevant content fragments matched to the request
Generate page layouts and copy using an LLM from source content
Retrieved source content and user context → LLM-composed page with generated layout and text
Serve tailored content responses to AI agents
AI agent request or query → Machine-consumable personalized content response
Adapt page structure per individual user instead of showing static pages
Individual visitor profile and behavior → Uniquely structured page for that specific user
Interpret user intent from incoming requests
Visitor's search terms, referrer, or interaction data → Inferred user intent used to drive composition
Why it made the leaderboard
If you're building web products in the agentic era, this frames a concrete shift: sites composed on the fly per visitor and structured for AI agents via WebMCP/A2A, with real economics (~1-2 seconds, a penny or two per page) attached to the idea.
Tags
Media
Comments (0)
No comments yet
Indexed by a proprietary survey. Corrections welcome.