Anthropic MCP Goes GA, Forcing Businesses to Rethink Agent Exposure
Anthropic's Model Context Protocol reached general availability this month, shifting MCP from experimental to infrastructure businesses must support. Combined with fresh citation research showing structured schema directly lifts AI recommendation rates, the cost of being agent-unreadable just got measurable.
This Week's Signals
Anthropic Declares MCP General Availability, Publishes Production Server Registry
Why it matters for your score
MCP moving to GA means Claude agents can now discover, authenticate, and transact with registered business endpoints in production — not just demos. Businesses that have not published an MCP server are invisible to any Claude-based agent workflow. The new public registry functions like a directory: if your business is not listed with a conformant server, you are not callable by the fastest-growing class of AI buyers.
BrightEdge Study Links JSON-LD ServiceSchema to 34% Higher AI Citation Rate
Why it matters for your score
BrightEdge's mid-2026 analysis of roughly 10,000 business pages found that sites with complete JSON-LD ServiceSchema and PriceSpecification markup were cited by ChatGPT, Perplexity, and Gemini 34% more often than equivalent pages without it. The effect was strongest for local service queries and SaaS category searches — exactly the queries where business owners want to be the recommendation. Schema is no longer a technical nicety; it is a ranking input for AI answer engines.
Perplexity Confirms PerplexityBot Now Prioritizes llms.txt Over Sitemap for Answer Sourcing
Why it matters for your score
Perplexity's crawler team confirmed in a July 2026 post that PerplexityBot's answer-sourcing pipeline now reads llms.txt first when present, using it to determine which pages carry authoritative, citable content before falling back to sitemap crawls. Businesses without an llms.txt file are letting the crawler make its own guesses about what to surface — frequently pulling outdated or low-signal pages. A well-structured llms.txt directly shapes which content becomes a cited answer.
What to do this week
- 1
Publish or update your llms.txt this week to reflect your highest-converting service pages and current pricing — use Faro's /tools/llms-txt generator to build a PerplexityBot-ready file that front-loads your authoritative content sections, given the confirmed crawler priority shift.
- 2
Run Faro's /tools/ai-schema audit on your top three service or product pages and add PriceSpecification and ServiceSchema JSON-LD where missing — the BrightEdge data shows this is the single highest-leverage structured-data change for lifting AI citation rates right now.
- 3
Register your business in Anthropic's MCP server registry by building a minimal conformant MCP endpoint — start with a read-only server exposing your product catalog, pricing, and availability, then use Faro's /tools/ai-readiness-scan to verify your endpoint, robots.txt crawler permissions, and llms.txt are all aligned before you submit.