Some brands are already attributing 10% of their revenue to AI agent-driven channels, from initial prompt to completed transaction. That claim comes from a Fortune commentary piece published today by a founder who says they’ve tracked nearly one billion agent interactions across agentic commerce platforms.

The timing is notable. McKinsey published research in October 2025 projecting that agentic commerce could drive up to $1 trillion in orchestrated US retail revenue by 2030, with global estimates reaching $3 trillion to $5 trillion. Six months later, the Fortune op-ed argues those projections are already playing out in early-adopter segments.

What the Numbers Show

The evidence trail across multiple sources points to agent-driven traffic becoming commercially meaningful:

  • Walmart: ChatGPT now accounts for roughly 20% of Walmart’s referral traffic, according to Modern Retail reporting on Similarweb data from September 2025. The Fortune article claims some retailers have seen that figure climb to 35%.
  • Etsy: ChatGPT drives more than 20% of referral traffic to Etsy, per the same Similarweb data reported by Modern Retail.
  • Target: ChatGPT referral traffic to Target is growing 40% month-over-month, according to Target’s own corporate communications from February 2026.

The Fortune piece argues that the discovery layer for commerce is now fundamentally different. Only 12% of URLs cited by AI tools overlap with Google’s top 10 results, according to Ahrefs research. A separate Semrush study found 90% of sources ChatGPT cited were not on Google’s first 20 pages.

The “Agent Experience” Argument

The op-ed’s central thesis is that brands now have two customer types: humans and agents. Traditional SEO optimization gets you in front of the first group. Getting in front of the second requires what the author calls “Agent Experience” (AX), a parallel to UX, focused on machine-readable content, structured product data, and agent-parseable FAQs.

One example cited: a robotics customer achieved a 94% increase in “agentic visibility” in four months by restructuring content away from human-readable marketing copy toward structured, specific answers that LLMs could extract and cite.

The practical playbook, according to the Fortune piece: audit how AI agents currently see your site, structure content for machine comprehension rather than just keyword ranking, monitor how your brand is referenced on Reddit and Wikipedia (which LLMs weight heavily), and build machine-readable product feeds via APIs and structured schemas.

What This Means for Agent Builders

The piece has a clear interest angle: the unnamed author runs a company in the agentic optimization space. That said, the underlying data from McKinsey, Similarweb, Ahrefs, and Semrush is independently verifiable.

For the agent ecosystem, the implications are concrete. If brands are adapting to serve agents as customers, agent platforms like OpenClaw become infrastructure for a new commerce layer. The question is no longer whether agents will participate in commerce, but how quickly the optimization ecosystem around agent-driven transactions matures.

The Fortune op-ed’s 10% revenue figure is self-reported by the author from their own client data, not independently audited. But the directional claim, that agents are already generating measurable commercial outcomes, is supported by the Similarweb referral data and Target’s own disclosure.