Half of All B2B Research Is Now Invisible to Your Attribution Stack

Your head of demand gen runs last month's report. Marketing-attributed pipeline is down 31%. Then sales walks in and tells you three enterprise accounts just closed. All three said they "did their research" before booking the call. Your website never saw them.

This is not an anomaly. It is the new baseline.

What the data says

As of March 2026, Similarweb data shows 50.2% of all search queries are now AI-assisted, up from 8% in September 2024. Eighteen months. One structural shift. In B2B specifically, the curve is steeper: AI-generated responses now appear in 61% of technology sector searches, 54% of financial services queries, and 49% of healthcare research.

Ahrefs reported that B2B SaaS companies with historically strong organic search performance saw average organic sessions drop 23% between Q1 2025 and Q1 2026. Brand search and direct traffic? Up 14%. Your buyers are finding you. More than before. You just cannot see how they got there.

The mechanism your stack cannot measure

Perplexity's Pro Business tier now has 340,000+ company subscriptions. Entire organizations, not individual researchers but entire buying teams, are conducting competitive analysis, vendor evaluation, and category research inside AI interfaces that do not generate a single trackable referral to the sources they are reading.

The dark funnel, previously estimated at 30-40% of B2B buyer touchpoints, has expanded to 60-70%. Forrester established that 70% of B2B buyers complete their research before making first contact with a vendor. That research has now relocated. Google searches with trackable clicks have given way to Perplexity sessions and AI Overviews that synthesize answers without sending traffic anywhere.

The result: your buyers arrive at the sales call more informed than at any point in recent history. They have read category comparisons, reviewed how your company stacks up against competitors, and formed a view of your credibility. None of it showed up in your attribution dashboard.

The wrong instinct

Most teams will respond by trying to fix attribution. Add a discovery question to every sales call. Buy a new tracking tool. Instrument the dark funnel better.

That is measuring the symptom. The structural question is different: are you in the sources AI is citing?

When a buyer queries "top B2B agencies for AI-native companies" in Perplexity, the response is built from citations. Those citations come from the publications AI systems treat as credible: Forbes, TechCrunch, VentureBeat, Reuters, HBR. Your website is not in that pool. A Fullintel-UConn study found that 89% of AI-cited links come from earned media. Muck Rack's analysis puts earned media at 82% of all AI citations, with 95% being non-paid content.

If you do not have editorial coverage in those publications, recent coverage with your actual positioning intact, you are not in the citations. You are not in the research brief your buyer generated this morning. You are invisible in 60-70% of the buyer journey you can no longer track.

Where this connects

The brands outperforming on the metrics that do still show up, brand search growth, deal velocity, enterprise win rates, share a common input: systematic editorial presence in the publications AI treats as ground truth. Research on earned media distribution shows that coverage across third-party news outlets increases AI citation rates by up to 325%, from roughly 8% citation share to 34% across major LLMs.

That gap explains why some brands appear in every AI-generated vendor brief in their category and others are never mentioned. It is not about ad spend or SEO rankings. It is about which brands have built the editorial record that AI models pull from when they synthesize recommendations.

You do not need better attribution tools. You need to be in the citations. That is the input worth fixing first.