Anthropic MCP Goes GA: Agent Commerce Infrastructure Is Now Table Stakes
Anthropic's Model Context Protocol reached general availability with expanded third-party server support, making structured agent-facing endpoints a baseline expectation rather than an experiment. Businesses without machine-readable service definitions are now actively invisible to the fastest-growing class of AI buyers.
This Week's Signals
Anthropic MCP Hits General Availability With Expanded Server Registry and OAuth 2.1 Auth Standard
Why it matters for your score
MCP GA means Claude-powered agents can now authenticate against and transact with registered business endpoints without human confirmation loops — the protocol has moved from developer toy to production infrastructure. Businesses with an MCP server listed in the registry are discoverable to agentic Claude sessions; those without one are skipped at the query-routing layer. If your service involves bookings, quotes, inventory, or purchasing, the absence of an MCP endpoint is now a structural gap your competitors can exploit.
Perplexity Sonar API Adds 'Source Freshness' Weighting, Favoring Sites With Crawlable, Timestamped Content
Why it matters for your score
Perplexity's Sonar retrieval layer now explicitly weights recency signals — meaning pages with valid dateModified schema markup and crawlable update timestamps rank higher in AI-cited answers. Sites that block PerplexityBot in robots.txt or lack Article/LocalBusiness schema with fresh dates are losing citation slots to competitors who've done the basic configuration work. This is a direct, confirmed ranking input, not a speculative best practice.
Google's 'AI Overviews' Begins Pulling Structured Pricing Data From Schema.org Offer Markup in Non-E-Commerce Categories
Why it matters for your score
Google confirmed that AI Overviews are now surfacing PriceSpecification and Offer schema from service businesses — agencies, consultants, SaaS tools — not just product retailers. If your pricing page has no structured data, Google's AI answer for 'how much does [your service] cost' will cite a competitor who does. This is the clearest signal yet that pricing transparency in machine-readable format is a direct AI discoverability input, not a nice-to-have.
What to do this week
- 1
Run Faro's AI Readiness Scan at /tools/ai-readiness-scan this week with specific attention to your robots.txt — confirm PerplexityBot and ClaudeBot are explicitly allowed, not just defaulting to a wildcard allow. The Perplexity Sonar freshness update means a misconfigured robots.txt is now a confirmed revenue leak, not a theoretical one.
- 2
Add or update PriceSpecification and Offer schema to your pricing or services pages using Faro's AI Schema tool at /tools/ai-schema — Google's AI Overviews confirmation means service businesses without this markup are handing citation slots to competitors. Include dateModified on all key pages while you're in the markup layer.
- 3
Generate or refresh your llms.txt file at /tools/llms-txt to include a structured MCP endpoint reference if you have one, or a placeholder services block if you don't — Anthropic's MCP GA means agent-facing sessions are now checking for machine-readable service definitions before routing queries. Pair this with Faro's OKF Generator at /tools/okf-generator to bundle your pricing, service scope, and availability into a format MCP-compatible agents can parse without a human in the loop.