Back to Market Signals
AI Toolshigh impact

AI Agent Protocols Like MCP and LLMs.txt Are Reshaping How Brands Get Found

By InnovaAI Research2 min read

New AI agent standards including MCP, A2A, and LLMs.txt are changing how brands surface in AI-driven search. Ahrefs data shows 97% of LLMs.txt files received no requests, exposing a gap between adoption and actual AI visibility.

Key Facts

01Ahrefs data shows 97% of LLMs.txt files received zero requests from AI crawlers, questioning current adoption value.
02New AI agent protocols including MCP, A2A, and ARD are actively reshaping how AI systems retrieve brand information.
03Anthropic now offers three distinct Claude model tiers (Sonnet 5, Sonnet 4.6, Opus 4.8) with different API pricing and performance tradeoffs.
04An identity gap exists when brands present inconsistent information across their website, search results, and AI-generated summaries.
05Structured data and schema markup now influence AI retrieval systems, not just traditional search engines.

Why It Matters

▶With 97% of LLMs.txt files getting no crawler requests, agencies recommending this tactic need accurate expectations to maintain client trust.
▶Clients with fragmented brand identities across search and AI surfaces are actively losing buyer confidence at the consideration stage.
▶Tiered AI model pricing means agency workflow costs can be reduced significantly by matching model capability to task complexity.
▶AI agent standards are being set now, and agencies that understand MCP and A2A early can offer advisory services competitors cannot.

Agency Actions

Search for your top five clients using ChatGPT, Perplexity, and Google AI Overviews, then document where AI-generated descriptions diverge from the client's own brand messaging.

low effort

Add LLMs.txt to client sites where practical, but set accurate expectations: current Ahrefs data shows 97% of these files receive no crawler traffic, so treat it as future-proofing rather than an immediate visibility driver.

low effort

Map your agency's AI model usage by task type, assigning lower-cost model tiers to drafting and summarizing tasks and reserving higher-capability models for complex reasoning or strategy work.

medium effort

Build an AI entity audit into your client onboarding process to identify how AI systems currently describe each client and where corrections are needed.

medium effort