Question 1

What is the AI dark funnel in B2B marketing?

Accepted Answer

The AI dark funnel refers to the portion of your B2B pipeline that was influenced by AI assistant recommendations — ChatGPT, Perplexity, Claude, Gemini — but leaves no referral trace in your analytics. When a prospect asks ChatGPT which CRM to evaluate and your company appears in the cited answer, that prospect may then Google your brand directly, navigate directly to your site, or click a LinkedIn ad — and every one of those touchpoints will be logged as branded search, direct, or paid. The original AI referral is invisible. In a Forrester survey from Q1 2026, 41% of B2B buyers reported using AI assistants as part of their vendor research process before making first contact. Of those, 73% said they did not click a link from the AI response — they separately searched for the vendor name. That behavior creates a structural dark funnel: AI-influenced demand that shows up in your analytics as organic branded, direct, and paid, making the AI source permanently invisible without deliberate measurement design.

Question 2

How do you measure revenue that came from AI search recommendations?

Accepted Answer

There is no single direct measurement method — AI assistants do not pass referral parameters, and most buyers do not disclose their discovery channel without being asked. The most reliable approach combines four proxy signals. First, track branded search volume lift in Google Search Console against your AEO citation rate improvement — a causal relationship is detectable over 60-90 days. Second, add a discovery question to every demo request, lead form, and sales call: ask explicitly whether the prospect used an AI assistant in their research. Third, correlate CRM pipeline velocity with AEO investment milestones — pipeline progression speed tends to increase for AI-influenced leads because they arrive with stronger pre-qualification. Fourth, run a quarterly attribution survey of closed-won deals, asking buyers to reconstruct their discovery journey. At scale, the four signals triangulate to a defensible pipeline estimate that most CFOs will accept as a directional attribution model, even without direct click-path data.

Question 3

Why doesn't GA4 show ChatGPT and Perplexity as traffic sources?

Accepted Answer

GA4 fails to capture most AI-referred traffic for two structural reasons. First, most AI assistant interactions do not generate a standard HTTP referrer header. When a ChatGPT user sees your company mentioned in an answer and then separately opens a new browser tab to search for your brand, the resulting session has no referrer from ChatGPT — it is classified as direct or organic search. Perplexity does send a referrer header on its inline citation clicks, which means perplexity.ai does appear in some GA4 reports, but only for the fraction of users who click the citation link directly rather than searching separately. Second, GA4's default channel grouping has no AI Search channel — Perplexity referrals that do pass a referrer header are bucketed into Referral or Unassigned. The fix requires a custom channel grouping rule in GA4 that captures known AI search domains: perplexity.ai, chat.openai.com, claude.ai, gemini.google.com, and copilot.microsoft.com. Even with this configuration, the majority of AI-influenced traffic remains invisible.

Question 4

What proxy metrics can you use to estimate AI search influence on pipeline?

Accepted Answer

Four proxy metrics together give a defensible estimate of AI search influence on B2B pipeline. Branded search volume trend is the most accessible: track your branded keyword impressions in Google Search Console on a rolling 28-day basis and correlate changes with your measured AI citation rate. A sustained uplift in branded impressions without a corresponding increase in paid brand spend strongly suggests AI-referred discovery. Direct traffic trend is the second signal — visits classified as direct that arrive on deep product or pricing pages (rather than the homepage) are disproportionately AI-influenced, because human direct traffic typically lands on the homepage while AI-primed prospects navigate directly to the pages the AI described. Demo-request form completion rate on first visit is the third: AI-influenced prospects arrive with a higher intent pre-qualification than cold organic visitors, producing a measurably higher same-session conversion rate. Fourth, average sales cycle length by cohort — AI-influenced pipeline closes 18-22% faster on average in the datasets we have tracked, because buyers arrive with vendor understanding already established.

Question 5

How should B2B CMOs report AI search attribution to leadership?

Accepted Answer

CMOs who try to report AI search as a direct revenue line lose the CFO argument immediately — the data does not support clean attribution. The framing that works in practice is a pipeline influence model rather than a last-touch revenue model. Build a dashboard with three components. First, AI citation share by category: what percentage of AI assistant responses to your category keywords include your brand, tracked weekly. This is the input metric. Second, branded search volume index: track branded impression volume as a proxy for AI-driven discovery, indexed against a pre-AEO baseline. Third, dark funnel pipeline estimate: take your average deal size, multiply by the estimated percentage of closed-won deals where AI research was confirmed via sales call discovery, and present this as a pipeline influence range with explicit assumptions. Frame the presentation as: here is how AI search is affecting the top of our funnel, here is how we are measuring it with the data we have, and here is the minimum investment required to increase the input metric by 20% in the next two quarters. That structure survives CFO scrutiny because it is honest about what is measurable and what is directional.