Goodie puts the AI-lead conversion multiplier at 2.1x. Competing vendors pitch 4x to 8x. Navattic pulled its own in-house data and found no lift at all, with signs its product-led leads from AI were worse. Same channel, same year, and the answer runs from “your best traffic source” to “a mild negative.”
A number that spans 8x to zero is measuring the person who quotes it.
Where do the conversion multipliers come from?
Vendor decks, with one exception. The 2.1x figure is Goodie’s, stated by its founder on the GEO Blueprint podcast, and Goodie sells an AI visibility platform. The 4x to 8x claims circulate from other vendors in the same market. Every one of these is a vendor claim: the company quoting the multiplier also sells you the channel. Note the market’s lowest number comes from a vendor too. Honesty grades on a curve here.
The measured exception is Webflow. Graphite, the agency running Webflow’s program, found LLM traffic converting 6x better than Google traffic, with 8% of Webflow signups arriving from LLMs (Ethan Smith on Lenny’s Podcast, September 2025). Smith has since widened the base: “I’m hearing the eight to ten percent number from multiple B2B engineering-forward companies,” he said at Surfer Academy. That is a signup-share figure across several companies, from an agency with its methodology on the table. Better evidence. An agency stat all the same.
Then there is first-party data from a company with no multiplier to sell. Surfer’s marketer, hosting that same Surfer Academy session, shared the company’s own funnel: 22% of self-attributed conversions point to ChatGPT or other AI, against 25% for Google search, “but Google sends us 30 times more traffic.” Per visitor, AI traffic converts at a different order of magnitude in Surfer’s funnel. Their method: a post-signup “how did you find us” survey with answer options in randomized order on every render.
Why can’t anyone agree on the number?
The middle of the funnel is untrackable, so every measurement method produces a different answer. Smith’s framing at Surfer Academy: the end of the funnel is the how-did-you-hear survey, the beginning is whether you appear in the answer, and “the in-between is very hard to track because there’s not that much clicking happening.” People read an AI answer, open a new tab, and search Google for the brand they just saw. Your analytics files that visit under direct or branded search.
Doola shows the size of the resulting gap. Over 25% of its customers say they came from an LLM. First-touch attribution shows single digits (GEO Blueprint podcast). Somewhere between the survey and the dashboard, two thirds of the customers evaporate.
Navattic sharpens the other edge. Around 15% of its inbound leads cite an LLM, per its growth lead on the same podcast. Its in-house conversion data shows no lift, and the PLG leads from AI may be worse. Volume arrived. Quality did not. Both facts sit in one funnel, which is why no universal multiplier exists. The figure depends on your product, your buyer, your price point, and which end of the broken pipeline you measured.
How do you test the multiplier on your own funnel?
Instrument both ends yourself and skip the middle. At the bottom: a how-did-you-hear-about-us field at signup, answer options randomized like Surfer’s survey, so position bias stops flattering whichever option sits first. At the top: repeated sampling of whether you appear when buyers ask their questions. Compare AI-attributed signups against search-attributed signups on your own conversion event. That quotient is your multiplier. It will not match anyone’s deck.
Our free scan runs the top end: 5 questions your buyers ask, answers and citations shown verbatim. The paid audit runs each question 7 times per engine, because a single AI answer is a coin flip, and every finding links to captured evidence. No ranks, ever. And no borrowed multipliers: if your funnel says 1.3x, that number beats a vendor’s 8x, because it is yours. Run the first pass free.
Sources
- Ethan Smith at Surfer Academy: the ROI of improving AI visibility
- The GEO Blueprint podcast: Goodie, Navattic, and Doola episodes (conversion and attribution figures as stated on air; vendor-sourced where noted)
- Lenny’s Podcast, Ethan Smith episode, September 2025 (Webflow 6x conversion and 8% signup share, Graphite-measured)