The Future of the Website: From Google-Centricity to an Interface Bundle for Agents

X postzeldmanOpinion

Jeffrey Zeldman highlights an essay by Joost de Valk analysing the existential shift facing websites. With Google no longer being the only surface worth optimising for, websites must evolve into interface bundles serving both human users and AI agents that complete actions directly.

Key points

  1. Websites have always served both humans and bots, but emerging AI agents lack Google's decades of experience parsing human-oriented page layouts.
  2. The emergence of GEO and AEO primarily reflects the need to optimize for multiple search and retrieval surfaces beyond just Google.
  3. Creating separate sites for bots (similar to past m-dot or AMP setups) leads to drift and failure; the solution is a single source of truth with multiple machine-friendly renderings.
  4. Websites are turning into bundles of interfaces, incorporating standards like MCP or command-line clients so agents can directly check availability or place orders.
  5. Conversion events are increasingly moving away from the website itself into external AI answers and conversational interfaces.

Why it matters

Web teams and business owners must rethink their technical architecture: relying purely on human-readable HTML and Google search ranking is no longer enough as AI agents become intermediaries for customer interactions.

Analysis

De Valk argues that technical excellence in web development is shifting from a rare expert skill to a selectable baseline managed via automated specifications. As AI tools handle the code mechanics and buying decisions occur off-site, a website's enduring value relies on having meaningful business content and actionable machine interfaces.

Original post

zeldman

@zeldman · 27 July 2026

“The work itself is the same as it always was. What changed is that Google is no longer the only surface worth the effort.” t.co/MHXvknTvmD

41 likes2 replies68 bookmarks9k views

View on X

Links and tools from this entry

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Similar in content, found by the AI search.

  • Link:Joost de Valk

    The Future of the Website: Interface Bundles and Verifiable Home Bases

    Joost de Valk explores the evolution of websites in the age of AI agents. Rather than disappearing, websites are shifting from human-only HTML pages into canonical systems of record. They must serve as interface bundles—including protocols like MCP—that allow automated agents to read data and take actions, while verifiable authenticity remains the only lasting competitive moat.

    AI & LLMs· Opinion

  • Link:Ross Hudgens

    Generative Engine Optimization: Upcoming AI SEO Book by Ross Hudgens

    Ross Hudgens, founder and CEO of Siege Media, has announced his upcoming book 'Generative Engine Optimization: The Definitive Guide to AI SEO'. It focuses on practical strategies to optimize content for AI answer engines such as ChatGPT, Google Gemini, Perplexity, and Claude.

    AI & LLMs· Resource

  • Article:Ross Hudgens

    Content and Conversation: Strategies for Generative Engine Optimization (GEO)

    On the Content and Conversation podcast, host Ross Hudgens and industry practitioners explore the shift toward Generative Engine Optimization (GEO). The discussions focus on AI search attribution, AI content quality, brand-level measurement in LLMs, and HubSpot's organizational pivot toward answer engine optimization.

    AI & LLMs· Discussion

  • Link:Ross Hudgens

    Generative Engine Optimization: Ross Hudgens' Guide to AI SEO

    Siege Media founder and CEO Ross Hudgens introduces 'Generative Engine Optimization: The Definitive Guide to AI SEO'. The book provides practical strategies to optimize digital content for AI answer engines, including ChatGPT, Google Gemini, Perplexity, and Claude, ensuring positive brand presence in LLM-generated answers.

    AI & LLMs· Resource

  • Video

    Video:AI Engineer

    Computer-Use Models Will Agentify the Web, Not APIs

    Dhruv Batra argues that AI agents will navigate the web via visual computer-use models rather than APIs or MCP servers. While high-traffic platforms might expose programmatic endpoints, the web's long tail of roughly 200 million active sites relies on legacy infrastructure, scanned files, and client-rendered states that only visual interaction can reliably parse.

    AI & LLMs· Talk

  • Link:Jeremy Howard

    The /llms.txt Specification v2 for Web-Browsing AI Agents

    Jeremy Howard presents version 2 of the /llms.txt proposal, specifying how websites can provide structured Markdown overviews and direct links so AI agents can navigate content efficiently.

    AI & LLMs· Resource

Let's talk about your project

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.