The receipt isn't the click. It's the shortlist appearance.
B2B buyers form the shortlist inside ChatGPT, Gemini, and Claude before they touch your site — and the click that used to credit the channel never happens. Crediting AEO with pipeline takes a new spine: shortlist appearance as the receipt, Position Score as the leading indicator, sales-call labeling as the attribution surface.
The Click Was the Receipt — Until It Wasn't
For two decades, B2B marketing attribution lived on a single artifact: the click. A buyer Googled, landed on a page, filled a form, became a contact, became a deal. Every dashboard — GA, HubSpot, Marketo, Salesforce — was built around that trail. If the click was tracked, the channel got credit.
That trail is breaking. B2B buyers now build their shortlist inside an AI answer — ChatGPT, Gemini, Claude, Perplexity — and most of them never click. They ask the model "what are the best tools for finance teams," read a paragraph naming three brands, and arrive at a sales call already short-listed. The decision happened. The click did not.
B2B Is the Segment Already Routed Through AI
It is tempting to file AEO under "interesting but premature." For consumer brands, that might still be defensible. For B2B, it is not. The medium has already moved, and the public data is unambiguous:
- 9 in 10 B2B buyers say AI has changed how they research vendors — not a future shift, a reported behavior change (G2).
- 1 in 2 buyers now start their evaluation with ChatGPT or a similar tool — a 71% jump in four months (G2).
- 55% of enterprise buyers rely on AI search more than Google (G2).
- Sales conversions from ChatGPT recommendations are up 436% (G2).
- 88% of enterprises use AI regularly, embedding it in the workflows where buying decisions form (McKinsey, State of AI).
The pattern is clear: the higher the deal value and the more research a buyer does, the more the model is doing the research. Enterprise buying — the longest cycle, the most peer-checking, the most "Google in a tab" — has flipped fastest. The buyer is still doing the work. They have moved the work into the model.
What the Model Is Reading When It Builds a B2B Shortlist
The reason B2B is so exposed to AEO is what the model retrieves from. When a buyer asks ChatGPT or Perplexity to recommend software, the model leans on a small set of third-party sources to decide who to name. For software queries, that distribution is extremely top-heavy:
| Source | Share of AI citations on software queries |
|---|---|
| G2.com | ~22% |
| TechRadar.com | ~17% |
| Zapier.com | ~16% |
| Salesforce.com | ~9% |
| HubSpot.com | ~8% |
| GitHub.com | ~7% |
| Gartner.com | ~7% |
| PCMag.com | ~5% |
| Capterra.com | ~5% |
| TechnologyAdvice.com | ~4% |
Source: G2 — "Do more G2 reviews mean more AI visibility." Data excludes Reddit, Wikipedia, and YouTube.
The implication for a B2B marketing team is direct. The AI answer your buyer sees is being assembled from a small, named set of sources. Your owned site is one input, and not necessarily the loudest one. If a rival has 400 G2 reviews and a complete profile and you have 40, the model has more evidence to retrieve about them than about you — and it will name them. Reviews, analyst coverage, and category sites are not vanity surfaces anymore. They are the corpus the model is summarizing from.
What Your Attribution Stack Stopped Seeing
The pipeline did not disappear. The visibility did. Demand routed through an AI answer leaves no UTM, no referrer, no landing page event. When a CFO opens GA and sees flat organic traffic, that is not evidence AEO is not working — it is evidence the analytics stack was instrumented for a medium that is now half the story.
Here is what an AEO-influenced deal actually looks like, end to end:
- A director at a Series C SaaS company asks ChatGPT: "best AI answer intelligence platform for enterprise marketing teams."
- The model names three vendors, including yours. Its evidence: a G2 profile, a TechRadar mention, and your own capability page.
- She does not click any citation. She copies the three names into a Slack: "anyone used these?"
- A colleague replies. Two days later she goes directly to your homepage and books a demo.
In the dashboard, that deal is "direct" or "dark social." Zero credit to AEO. Zero credit to the answer that put you on the list. The model did the most important thing a marketing channel can do — get you on the shortlist before the buyer ever saw your site — and the attribution stack reports nothing.
Five Signals That Survive the Click Going Dark
Attribution in an AI-mediated market does not vanish — it moves. The signals are still measurable; they are no longer in web analytics. A team running AEO seriously instruments these five, and reports on them as deliberately as the old funnel:
1. Direct sales-call language
The cleanest signal is the buyer telling you. "We saw you in ChatGPT," "an AI tool recommended you," "you came up when I asked about X" — phrases that did not appear in discovery calls a year ago. Instrument them: a single Salesforce field on the discovery template — Heard about us via AI? Y / N, which model, which query if known — produces a labeled stream of AEO-influenced pipeline within a quarter.
2. Position Score and Answer Share
The leading indicator. Track where the brand is named — by buyer, by use case, by rival — across the five major models, and watch the trajectory. A Position Score that climbs from 38 to 64 over a quarter is not a vanity metric; it is the shortlist appearance rate compounding. If the score is moving up and self-reported AI-influenced deals are following, the chain is closed.
3. Branded direct traffic and search lift
The downstream tell. When the model recommends a brand by name and the buyer does not click the citation, they often type the brand into the URL bar or Google it to verify. Branded direct and branded search lift correlate tightly with answer-share gains — not 1:1, but tightly enough to use as a reasonableness check on the Position Score story.
4. Inbound demo and contact-form quality
An AEO-influenced lead arrives further down the funnel than a cold one — they already know what you do, they already named you against rivals, they want to talk pricing or fit. Form-to-meeting conversion, sales-cycle length, and average deal size on "direct" inbound should all improve as AEO ramps. If they are, the dark pipeline is real.
5. "Where did you first hear about us" — asked properly
The oldest tool in marketing, used badly by most teams. Add ChatGPT / Claude / Gemini / Perplexity / other AI tool to the closed-won survey and the demo-booking form. Within two quarters the share of buyers crediting an AI tool will tell you what the attribution stack cannot.
The CFO Conversation
A marketing leader asked to defend AEO budget to a CFO has three pieces to put on the table.
One: the medium moved, and the data is public. Half of enterprise buyers start research in ChatGPT. Nine in ten say AI changed how they research. This is not a thesis; it is a reported behavior shift, sourced from G2 and McKinsey. The question for the CFO is not "should we invest in AEO." It is "are we comfortable being absent from the surface where half the buyers are now forming their shortlist."
Two: the leading indicator is the Position Score, not last quarter's pipeline. AEO is a compounding asset, not a paid channel. Pipeline lags the position move by a quarter or two. The right metric to bring to a CFO is the position trajectory across rivals, by buyer and use case, with a stated cadence of monthly review. A CFO can defend an investment with a leading indicator; they cannot defend one with hope.
Three: the alternative is invisible compounding loss. Every quarter a rival is named in answers your brand should appear in, the model retrieves their proof on the next adjacent query, and the next, and the next. AEO position stacks. A team that delays loses ground that is increasingly expensive to retake. The cost of inaction is not zero; it is the compounding gift to whichever rival is working the surface now.
| Pre-AEO attribution | AEO attribution |
|---|---|
| Click is the receipt | Shortlist appearance is the receipt |
| GA / HubSpot funnel | Position Score + sales-call labeling + branded lift |
| Channel credited at the click | Channel credited at the recommendation |
| Pipeline visible in real time | Pipeline visible at discovery + closed-won |
What Changes When You Instrument It
Three things change once a team measures AEO the way it actually moves pipeline.
- The dark pipeline becomes a number. "Direct" traffic stops being a black box. A measurable share gets reattributed to the answer that put you in the consideration set, and the marketing team can defend that share with the sales-call data behind it.
- Budget moves toward what compounds. Once Position Score gains correlate to inbound quality and self-reported AI-influenced deals, the case for moving spend from one-shot campaigns to proof that stacks across answers — review velocity, capability evidence, analyst coverage, citable benchmarks — writes itself.
- The marketer's defense changes. A leader who can point to a Position Score climbing, a sales-call field filling with AI mentions, and inbound quality improving has the same defensibility as a PPC manager with a CAC chart — but on the surface that is replacing search.
The Standard to Hold
A marketing team running AEO seriously should be able to answer three questions for a CFO at any moment: where is our position moving, how much pipeline is the sales team labeling as AI-influenced, and what is it costing the business when a rival is in the answer and we are not. If the answer to any of those is "we don't track it," the team is still attributing to a medium that no longer carries the deal.
The click is not coming back as the receipt. The shortlist is. The teams that get there first — instrumenting the signals that survive, defending the budget with leading indicators, naming the cost of absence — will route the pipeline that used to flow through search through the answer engines instead. The medium changed. The receipts changed with it.
