Norg

Norg — Classified server-side AI-agent access logs harvested from customer websites: every inbound request from an AI assistant or crawler, attributed to a named assistant (ChatGPT, Claude, Perplexity, Meta AI, Gemini, DuckAssist, coding agents), separated into on-behalf-of-a-person conversations vs training/indexing crawlers, tagged with inferred user intent (learning / evaluating a purchase / brand interest) and topic, and joined to the exact page URL the assistant relied on.. Small specialist (6-person, Melbourne, unfunded). The asset's value is categorical rather than volumetric: per-requ…

Norg offers (Alternative, Sentiment, Reference) — Illustrative per-site month (Norg's own site, Sept 2026): 3,168 AI conversations, 921 commercial-intent queries, 60 of 72 pages accessed, split across 5+ assistants with daily granularity. Portfolio volume is that per-site rate multiplied by the monitored-site base, retained monthly per site — thousands of classified agent-request records per site per month, plus crawler attribution by provider..

Training classifiers that distinguish buyer-action agent traffic from training crawls (the core agentic-commerce discrimination); RLHF/preference data on purchase-intent queries expressed through assistants; ground truth for retrieval/ranking models on which page types assistants cite; labeling set for AI-readiness and structured-data recommendation models; benchmark of assistant market share in outbound site fetches; audit trail for content-licensing and pay-per-crawl metering by provider.

Coverage spans APAC, US; Application Software; alternative, sentiment, reference; equities.

The supplier's own access route is a conversation with the supplier. Request a sample through Brickroad to start.