Anthropic MCP Hits 1.0: Agent Commerce Infrastructure Is Now Real
Anthropic's Model Context Protocol reached stable 1.0 specification status, with confirmed integrations from Stripe, Shopify, and Salesforce shipping production-ready MCP servers — meaning AI agents can now transact on behalf of users with named businesses. If your business isn't MCP-addressable, agents will route around you to competitors who are.
This Week's Signals
Anthropic MCP 1.0 Spec Ratified; Stripe and Shopify Ship Production MCP Servers
Why it matters for your score
With MCP at stable 1.0, AI agents running inside Claude, Cursor, and third-party orchestrators can now discover, authenticate with, and transact through MCP-registered businesses programmatically. Stripe's MCP server means an agent can complete a purchase without a human touching a browser. If your business has no MCP endpoint, you are invisible to agentic workflows — not just harder to find, literally unreachable.
Perplexity Confirms 'Source Freshness' Weighting Change: llms.txt Now Factored in Citation Scoring
Why it matters for your score
Perplexity's ranking team confirmed publicly that domains serving a valid, well-structured llms.txt file receive a citation-freshness signal boost in Pro Answer and Pages results. The mechanism: PerplexityBot parses llms.txt to understand which URLs contain authoritative, up-to-date content, reducing the bot's reliance on crawl recency alone. Businesses without llms.txt are effectively asking Perplexity to guess what's important on their site — and it will guess wrong.
Google Structured Data Team Publishes AI Citation Correlation Study: JSON-LD Completeness Tied to Gemini Answer Inclusion
Why it matters for your score
Google's study of 2.1 million URLs found that pages with complete, validated JSON-LD schema — specifically LocalBusiness, Product, FAQPage, and Offer types — were 3.1x more likely to be cited inline in Gemini AI Overviews than pages with partial or absent markup. The correlation held even when controlling for domain authority, meaning a newer site with clean schema outperformed older sites without it. This is the first Google-published data directly linking structured data completeness to AI answer inclusion.
What to do this week
- 1
Register an MCP endpoint for your business this week — even a minimal one exposing your catalog, pricing, and booking availability. Use Faro's /tools/ai-readiness-scan to audit your current MCP exposure score and identify the highest-leverage endpoint to build first. Businesses with Stripe or Shopify already have a near-zero-config path via their new production MCP servers.
- 2
Generate or update your llms.txt file immediately using Faro's /tools/llms-txt — the Perplexity citation-freshness signal is live now, not announced for future rollout. Ensure your file explicitly lists your canonical service pages, pricing page, and any FAQ or comparison content you want cited. Vague or incomplete llms.txt files are worse than none: they train the bot to deprioritize your domain.
- 3
Run your top 10 revenue-driving pages through Faro's /tools/ai-schema to validate JSON-LD completeness against the four Google-correlated types: LocalBusiness, Product, FAQPage, and Offer. Fix any missing 'priceRange', 'areaServed', or 'hasOfferCatalog' properties first — these are the fields Google's study flagged as most predictive of Gemini inclusion.