TL;DR – Key Takeaways
- AI search now drives a quarter of our new revenue. ChatGPT and Claude together influenced 25% of our net new MRR in June – more than double Google’s 11%.
- Claude is our single best-converting channel. Signups who discovered us via Claude convert to paid at 9.5% – nearly twice Google’s 5.0% and more than double ChatGPT’s 4.2%.
- Claude users punch above their weight. They’re only 10% of signups but influence 18% of revenue – and the average Google-sourced customer spends 15% less than the average Claude-sourced customer.
The takeaway: AI search isn’t just a traffic story anymore. It’s a revenue story. That’s why we’re doubling down on both GEO and SEO.
This is the second post in a series on how we measure AI search at OtterlyAI. In our first analysis, we compared self-reported attribution against last-touch attribution in Google Analytics 4- and explained why there’s a gap between the two. This post goes one level deeper: revenue. Which channels actually influenced our net new Monthly Recurring Revenue (MRR) in June, and how the answer changes depending on which data you trust.
Net new MRR is the recurring revenue added in a period from new customers. It’s the number that tells you whether the business is actually growing through net new business.
Which channels influenced our revenue?
Looking at our self-reported attribution report for June – where every new signup tells us, in their own words, where they first heard of us – the ranking is clear:
- Claude influenced 18% of our net new MRR
- Google influenced 11%
- ChatGPT influenced 7%
- All other reported channels combined (direct, word of mouth, email, and so on) account for the remaining 63%
Add Claude and ChatGPT together and AI search influenced 25% of our net new MRR — more than double what Google contributed.
Claude has the best Conversion Rate to Paid
The revenue share alone doesn’t tell the whole story. When we look at how signups from each channel convert to paying customers, Claude wins by a wide margin:
| Channel | Signup → paid conversion |
|---|---|
| Claude | 9.5% |
| 5.0% | |
| ChatGPT | 4.2% |
The paradox: Google sends signups, Claude sends customers
Here’s the part I find fascinating. Only 10% of new signups told us they came from Claude — yet those users influenced 18% of revenue. Compare that to Google, which drove 16% of signups but influenced only 11% of revenue.
Google sends us more signups. Claude sends us more customers.
Put the whole funnel side by side and the pattern is obvious:
Claude users also spend more
Claude-sourced users don’t just convert better than users from Google or ChatGPT — they also spend more once they’re customers. The average customer who found us via Google spends 15% less than the average customer who found us via Claude. (See chart below.)
Why we trust self-reported attribution over last-touch
Self-reported attribution has real flaws, and I won’t pretend otherwise. Is what a user picks in a survey always accurate? No. Some people skip ahead and click whatever’s first. Fair.
But last-touch attribution and web analy tics are flawed too — arguably more so for top-of-funnel channels. Brand marketing rarely shows up cleanly in analytics the way branded, direct, or organic traffic does, because those are bottom-of-funnel, more “direct” channels. The demand that AI search creates tends to resurface later as “direct” or “word of mouth.”
In my last role as VP of Marketing, I spent millions on brand marketing — PR, social, marketing stunts. I never measured those channels by last-touch traffic or revenue. I measured them by two things: brand-visibility metrics and self-reported attribution. Did customers see the stunt? Did they tell us — via a form or our sales team — that they saw the billboard or heard the radio spot?
That’s the same logic we apply to AI search today. And by that logic, ChatGPT and Claude have quietly become two of our most valuable acquisition channels.
How to measure AI search revenue for your own company
If you want to run this analysis yourself, here’s the repeatable version of what we do:
- Add a self-reported attribution field at signup. One open or semi-structured question: “How did you first hear about us?” This is your single most important AI-search measurement tool, because AI referrals rarely survive in analytics.
- Tag AI-search sources explicitly. Bucket answers into ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews — not a generic “AI” catch-all. The engine-level differences (see our 9.5% vs 4.2%) are where the strategy lives.
- Tie the field to revenue, not signups. Join the attribution answer to the customer’s plan and MRR in your CRM. Conversion rate and revenue share reveal what signup counts hide.
- Compare self-reported against last-touch. Run both side by side. A large, consistent gap in favor of self-reported is your evidence that AI search is being undercounted.
- Track it monthly, alongside brand-visibility data. Attribution tells you whether AI search converts; AI-search visibility monitoring tells you why — which prompts you show up for, and where you don’t.
Common pitfall: treating AI search as one channel. ChatGPT and Claude behave nothing alike in our data — different volume, different conversion, different customer value. Measure them separately.
What this means for your growth strategy
Strip out the specifics and three durable lessons remain:
- People remember AI search. When we ask new signups where they discovered us, ChatGPT and Claude come up as primary channels — unprompted. That’s brand recall, and it’s the hardest thing in marketing to buy.
- AI search converts. Lead-to-paid conversion from Claude beats every other channel we track, including Google. High-intent traffic doesn’t need to be re-sold.
- AI-search customers are worth more. Users from Claude and ChatGPT convert at a higher rate and spend more than users from Google.
Traffic that remembers you, converts better, and spends more isn’t a channel you experiment with — it’s a channel you invest in. Which is exactly why we’ll keep doubling down on GEO (Generative Engine Optimization) alongside SEO.
Try OtterlyAI free for 7 days.
Track your brand’s visibility across every major AI engine, see which prompts you show up for, and start turning AI search into revenue – no credit card required. Start your 7-day free trial →
How we measure this
Attribution: We look at two data sources side by side — self-reported attribution (every new product signup is asked where they first heard about us) and Original Traffic Source (from CRM reporting).
