Can lead sources be tracked, and how do we tell whether a B2B inquiry came from AI search (ChatGPT/Perplexity/Gemini) versus SEO, ads, or social?

• Trackable with high confidence: session-level source signals, including UTM parameters, landing page URL, and HTTP referrer (referer) when available.
• Not guaranteed: identifying the exact AI-generated answer that mentioned your company, because many AI tools may not pass referrer data consistently, and users may copy/paste links or open in new contexts.

Goal: make every channel land on a measurable URL, and store it with the inquiry.

• Create channel-specific URLs using UTM, e.g. utm_source=ai, utm_source=seo, utm_source=linkedin.
• Use channel-specific landing pages where practical (e.g., /ai/faq/ vs /seo/solutions/) to reduce ambiguity.
• On inquiry forms, store the following in hidden fields:
  • utm_source / utm_medium / utm_campaign
  • first_landing_url (initial landing URL)
  • current_page_url (form submit page)

Result: when a lead is created, the CRM record contains a reproducible path: Channel → Landing → Inquiry.

Goal: distinguish AI-driven sessions when referrer data is missing.

• Use a dedicated short link for AI distribution (e.g., shared in AI-facing knowledge assets):
  • Short link resolves to a URL with utm_source=ai
  • Short link logs click timestamp and destination URL
• Create an AI-only quotation request page or AI-only datasheet download page (gated), so that AI-origin intent is captured with a unique entry point.

Result: even if the browser referrer is empty, the utm_source=ai plus AI-only landing path provides deterministic classification in most cases.

Goal: ensure consistent attribution across sales, marketing, and reporting.

ABKE/AB客 rule set (example):

1. Classify as “AI” if:
   • UTM indicates AI (e.g., utm_source=ai) OR referrer domain is detected as an AI-related domain (when available),
   • AND the first landing page is a technical FAQ/specification page that contains decision-critical fields such as:
     • Model / SKU
     • MOQ
     • Incoterms (e.g., EXW, FOB, CIF)
2. Otherwise, classify by standard rules:
   • SEO: organic search UTM/referrer + informational/product pages
   • Ads: paid UTM (e.g., utm_medium=cpc)
   • Social: social UTM/referrer (e.g., utm_source=linkedin)

Result: attribution becomes repeatable and auditable, not dependent on subjective sales notes.

• Lead record fields: utm_source, utm_medium, utm_campaign, first_landing_url, current_page_url
• Web analytics: session source/medium aligned with CRM record
• Click logs (if short links used): short link ID → destination URL → timestamp match

• Referrer may be blank: some AI apps/browsers open links without passing referrer information; classification then relies on UTM and landing patterns.
• Copy/paste behavior: if a user copies a URL and shares it internally, attribution may shift unless UTM is preserved.
• Not “answer-level” traceability: you can usually identify the session and the path, but not always the precise AI response that triggered the click.

In B2B sourcing, AI-origin visits often land directly on decision pages (FAQ/specs/terms) rather than broad marketing pages. Capturing model, MOQ, and Incoterms on those landing pages improves both source attribution and sales qualification in the CRM.