Faro AI Signals·July 5, 2026

Anthropic MCP Goes GA, Forcing Agent Discoverability Into Boardrooms

Anthropic's Model Context Protocol reached general availability with a wave of third-party integrations, making MCP exposure a concrete business requirement rather than a speculative bet. Simultaneously, llms.txt adoption crossed measurable thresholds among e-commerce and SaaS platforms, and new citation research confirms structured schema directly influences which businesses Claude and Perplexity surface in recommendations.

ShareShare

This Week's Signals

Anthropic MCP Reaches General Availability With Shopify, Stripe, and Atlassian as Launch Partners

Source: Anthropic ↗

Why it matters for your score

MCP GA means Claude agents can now natively query your business's inventory, pricing, and service catalog in real time — but only if you have an MCP-compatible endpoint exposed. Shopify and Stripe merchants with MCP connectors configured are being surfaced directly inside Claude's agentic shopping and payments flows, bypassing traditional search entirely. If your business lacks an agent-facing API or MCP server, you are structurally invisible to this growing class of autonomous purchasing agents.

Perplexity Business Recommendations Now Explicitly Weight JSON-LD LocalBusiness and Product Schema in Citation Ranking

Source: Search Engine Land ↗

Why it matters for your score

An analysis of Perplexity's answer engine behavior confirmed that businesses with complete JSON-LD Product, LocalBusiness, and Review schema were cited 3.4x more frequently than equivalent businesses without it — controlling for domain authority. This is the clearest signal yet that structured data is functioning as an AI citation signal, not just a traditional SEO signal. Businesses that haven't audited their schema for AI-consumption completeness — including priceRange, areaServed, and hasOfferCatalog fields — are leaving recommendation slots on the table.

llms.txt Adoption Doubles Among SaaS and E-Commerce Domains as Cloudflare Adds Native Support

Source: The Verge ↗

Why it matters for your score

Cloudflare's Workers and Pages platform shipped native llms.txt generation and hosting as a one-click feature, dropping the implementation barrier to near zero for the roughly 20% of the web running on Cloudflare infrastructure. Measured crawl data from CommonCrawl's June 2026 index shows llms.txt presence has doubled on SaaS and e-commerce domains in 90 days. AI crawlers including ClaudeBot and PerplexityBot have been confirmed to fetch llms.txt before crawling deeper pages — meaning your llms.txt is now the first editorial statement your business makes to an AI agent.

What to do this week

1
Audit your MCP exposure this week using Faro's AI Readiness Scan (/tools/ai-readiness-scan) — specifically check whether your business has a discoverable MCP server or agent-facing API endpoint. If you're on Shopify or Stripe, prioritize configuring the MCP connectors announced at GA; these are live recommendation surfaces inside Claude agents now, not a future roadmap item.
2
Run your site through Faro's AI Schema tool (/tools/ai-schema) and verify your JSON-LD includes priceRange, areaServed, hasOfferCatalog, and aggregateRating fields — the exact properties the Perplexity citation ranking analysis flagged as decisive. Incomplete schema is now a measurable revenue leak, not a technical nicety.
3
If you haven't published an llms.txt file, generate one now via Faro (/tools/llms-txt) before Cloudflare's native tooling makes your absence conspicuous by comparison. ClaudeBot and PerplexityBot are confirmed to fetch it on first crawl — your llms.txt controls the summary an AI agent reads about your business before it reads anything else on your site. Treat it like your above-the-fold homepage copy, not a robots.txt afterthought.

Sources

← All editions Scan your site now →