
SEO agency
95% clarity about marketing KPIs and a foundation to scale
Nikita Yatsun
Managing Director
Read the storyUse case
GA4's channel groups were defined before AI search existed. LeadJourney reports every engine as its own channel, with the revenue each one produced.
Rated by the teams who fixed this first
One visit, before and after
Same visit, same form fill. The difference is a tracking script that recognises AI referrals the moment they arrive, instead of filing them under direct.
The problem
When ChatGPT or Perplexity recommends you, the recommendation happens inside the engine, where no analytics tool can look. What arrives on your site is a visit with nothing to classify it, so the fastest growing channel in B2B reports as nothing at all.
GA4 is classifying with pre-AI categories
Every engine recognised on arrival
The fix
Two places, one answer. Marketing sees every engine as a channel in LeadJourney, with the same lead and revenue columns as Meta and Google. Sales and the budget meeting get the numbers in the tools they already read.
The dashboard where AI search becomes a channel
Where sales and the budget meeting look
Watch it
The four AI engines separated out of direct traffic, followed through to meetings and revenue, with no code.
What changes
Every one of these is a report in LeadJourney, not a data project. The engines keep sending traffic either way; the difference is whether you can see it.
How many leads did ChatGPT send us last month?
Which AI engine produces leads that actually close, not just visits?
How much pipeline and won revenue does AI search create compared to paid?
Which pages do the AI engines send their visitors to?
Is our AI search traffic growing faster than classic organic?
Which of our closed deals had an AI answer as the first touch?
Case studies
One measured result per customer, reported by the customer. The stories closest to this use case come first.

SEO agency
95% clarity about marketing KPIs and a foundation to scale
Nikita Yatsun
Managing Director
Read the story
LinkedIn growth agency
2x customer retention, once every LinkedIn touchpoint was tracked to revenue
Marc Richard
CEO
Read the story
Tax and accounting
54% better lead quality after using LeadJourney for marketing
Karsten Guhr
Managing Director
Read the story
Klickkraft GmbH
Business coaching
€10-20k in monthly Meta spend, tied to qualified leads and revenue per ad
Florian Buck
Managing Director
Read the story
![]()
Digital publishing
Which article produced the lead and the revenue, across AI and organic search
Andreas Malkow
CEO & Head of PR
Read the story
SEO and GEO agency
Proven ROI on every retainer, from organic and AI search leads through to revenue
Philipp Nessmann
Managing Director
Read the storyGetreach
Backlinks marketplace (SaaS)
Setup
No channel definitions to maintain, no regex filters, nothing to tag.
FAQ
If yours is not here, our team answers in the chat within a few minutes.
GA4 classifies sessions with channel definitions written before AI search existed, so an AI referral has no channel of its own to land in. Clicks out of the ChatGPT app often arrive without a referrer at all, and a visit without a referrer is direct by definition. It is not a GA4 bug, it is a category that did not exist when the categories were made.
ChatGPT, Perplexity, Claude, Gemini, Copilot, Grok and DeepSeek referrals are reported as their own sources instead of landing in direct. There is nothing to configure: the tracking script separates them on its own, and new engines are added as they appear.
No. AI search tracking is on by default. The same script that records your sessions recognises AI referrals on arrival, so there are no channel groups to define, no referrer lists to maintain and nothing to tag.
Yes, and that is the point of doing this in an attribution platform rather than an analytics filter. Every AI visit joins a journey, the journey is stitched to the lead when the form is submitted, and the lead is matched to pipeline and closed revenue in your CRM. So each engine reports in leads, pipeline and won revenue, the same columns as your paid channels.
No tool can attribute a recommendation nobody clicked, and we will not pretend otherwise. What LeadJourney does measure is every visit that does happen: the click out of the answer, and the journey after it, so a lead who returns later through brand search still credits the AI touch that started the journey.
You can build custom channel groups for the referrers that arrive labelled, and you will still miss the app clicks that carry no referrer and everything the browser-side script loses to ad blockers and iOS. LeadJourney tracks server-side at 95%+ accuracy and reports leads and revenue rather than sessions, which is the number the budget meeting actually asks for.
In most analytics setups, yes, because Gemini sits inside Google. LeadJourney splits Gemini out from Google organic and attributes it on its own, so your AI search column and your SEO column stop contaminating each other.
Yes. Data is processed and stored in the EU, we sign a data processing agreement, and tracking runs first-party and server-side, so it respects consent decisions rather than working around them.
AI search, measured like paid
Book a demo and watch your own direct traffic split into ChatGPT, Perplexity, Gemini and Claude, live on your funnel.