Anthropic Embeds MCP Into Claude Enterprise, Forcing Business Readiness Now
Anthropic's full MCP integration into Claude for Enterprise — confirmed shipping in Q2 2026 — means AI agents are now actively querying structured business endpoints, not just crawling pages. Businesses without agent-facing APIs or llms.txt files are invisible to a growing share of agentic workflows.
This Week's Signals
Anthropic Confirms MCP Is Default Infrastructure in Claude for Enterprise
Why it matters for your score
Claude for Enterprise now routes agentic tasks through MCP-compatible endpoints by default, meaning businesses with registered MCP servers get queried directly while others are bypassed entirely. This is no longer an experimental protocol — it is the delivery layer for enterprise AI purchasing and vendor discovery. If your business doesn't expose an MCP endpoint, Claude agents handling procurement or vendor research will not surface you.
Perplexity Expands Structured Citation Weighting, Penalizes Unstructured Pages
Why it matters for your score
Perplexity confirmed an update to its answer engine ranking logic that explicitly upweights pages with valid JSON-LD schema — specifically LocalBusiness, Product, Service, and FAQPage types — when generating cited responses. Pages without structured markup are being demoted in citation selection even when they rank well in traditional search. For any business whose customers ask Perplexity for recommendations, missing schema is now a direct revenue leak.
llms.txt Adoption Crosses 40,000 Verified Domains as Tooling Ecosystem Matures
Why it matters for your score
The llmstxt.org directory crossed 40,000 verified adopters in June 2026, with the fastest growth among SaaS, professional services, and local business categories — exactly the segments AI agents are queried about most. More critically, GPTBot and ClaudeBot crawl logs analyzed by third-party researchers show both crawlers now preferentially parse llms.txt before crawling full site content, meaning your llms.txt file is often the first — and sometimes only — document an AI reads about your business. An absent or malformed file means an AI agent forms its understanding of you from whatever fragmented content it finds.
What to do this week
- 1
Run your site through Faro's AI Readiness Scan (/tools/ai-readiness-scan) and filter specifically for MCP endpoint status and llms.txt validation — given Anthropic's enterprise rollout, these two signals are now the highest-priority gaps to close before Q3 planning cycles when enterprise procurement agents are most active.
- 2
Use Faro's AI Schema tool (/tools/ai-schema) to add or audit JSON-LD markup for Service, LocalBusiness, or Product types — Perplexity's confirmed citation weighting update means valid structured data is now a direct input to whether your business gets cited in answer engine responses, not just a technical best practice.
- 3
Generate or refresh your llms.txt file via Faro's llms.txt generator (/tools/llms-txt), ensuring it includes your core service descriptions, pricing signals, geographic coverage, and a clear differentiation statement — with both GPTBot and ClaudeBot confirmed to parse this file early in crawl sequences, it is now your primary first impression with AI agents, not your homepage.