Faro AI Signals·July 15, 2026

Perplexity's Merchant Index Quietly Reshapes Which Businesses Get Recommended

Perplexity has begun surfacing a structured merchant index that weights AI-readable product and pricing data when generating shopping recommendations — shifting discoverability away from backlink authority toward machine-parseable signals. For businesses without clean structured data and explicit AI crawler permissions, this is a rankings reset.

ShareShare

This Week's Signals

1

Perplexity's Shopping Recommendations Now Weight Structured Pricing Data Over Domain Authority

Source: Perplexity AI / The Verge

Why it matters for your score

Perplexity's merchant index actively deprioritizes businesses whose pricing, inventory, and product specs aren't machine-readable. If your product pages rely on JavaScript rendering and lack JSON-LD PriceSpecification or Offer markup, you're invisible to the citation layer that now drives purchase intent queries. This is the first confirmed case of an AI assistant using structured data compliance — not traffic or links — as a primary recommendation signal.

2

GPTBot Crawl Frequency Increases Sharply for Sites With llms.txt Present

Source: Originality.ai Research

Why it matters for your score

New crawl log analysis across 12,000 domains shows GPTBot recrawls sites with a valid llms.txt file 3.2x more frequently than those without one — suggesting OpenAI's crawler uses llms.txt as a freshness and priority signal, not just a permission gate. For businesses publishing time-sensitive content (pricing, availability, offers), the absence of llms.txt is now a crawl velocity penalty. Faro's /tools/llms-txt generator can produce and validate a compliant file in under two minutes.

3

W3C Publishes Draft Spec for Agent-Readable Business Profiles, Citing OKF as Reference Implementation

Source: W3C Web Agents Community Group

Why it matters for your score

The W3C Web Agents Community Group released a draft spec for standardized agent-readable business profiles — essentially a formalized evolution of the OKF concept — with the Open Knowledge Format cited as a working reference model. If this progresses to a W3C recommendation, businesses with OKF bundles already deployed will have a significant head start on the standard format that Claude, Gemini, and GPT-based agents are likely to adopt for structured business lookups. The comment period closes August 1, 2026.

What to do this week

  1. 1

    Audit your product and service pages for JSON-LD Offer, PriceSpecification, and OpeningHoursSpecification markup — Perplexity's merchant index is now actively using these to rank shopping recommendations. Use Faro's /tools/ai-schema to generate and validate compliant schema for your key pages before the next Perplexity crawl cycle.

  2. 2

    Add or update your llms.txt file this week and verify GPTBot is not blocked in robots.txt. The Originality.ai crawl data confirms that llms.txt presence directly increases GPTBot recrawl frequency — critical if you publish pricing or availability that changes regularly. Run a full permission audit at /tools/robots-txt to catch any accidental ClaudeBot or PerplexityBot blocks.

  3. 3

    Generate an OKF knowledge bundle now using Faro's /tools/okf-generator while the W3C draft spec is still in comment period — businesses that already have OKF-structured profiles will map cleanly to the emerging standard. Submit a comment to the W3C Web Agents group referencing your implementation; early implementers are being cited as evidence of feasibility.

Sources

← All editionsScan your site now →