Tax and accounting
54% better lead quality after using LeadJourney for marketing
Karsten Guhr
Managing Director
Read the storyAI search engine tracking
AI search engine tracking that reports every engine, not "Direct".
ChatGPT, Perplexity, Gemini, Claude, Copilot, Grok and DeepSeek, each named on arrival and followed to the deal in your CRM.
- Seven engines, nothing to configure
- Recorded first-party, server-side
- Leads and revenue per engine, from your CRM
Rated by teams that stopped guessing what ChatGPT sends
- Trustpilot4.9 out of 5 on Trustpilot
- G25.0 out of 5 on G2
- Capterra5.0 out of 5 on Capterra
AI search tracking for marketing teams at
Every engine, last 30 days
Your workspace
AI search enginesEvery engine as its own source, revenue from the CRM, last 30 days
ChannelVisitsLeadsRevenue
ChatGPT1,84051€142.0k
Perplexity1,12029€78.4k
Gemini64016€41.2k
Claude3889€24.6k
Copilot4127€18.8kAI search, all engines, 112 leads€305.0k
GA4 reported 880 of these 4,400 visits with a usable referrer. The rest sat in Direct and Unassigned.
The definition
What is AI search engine tracking?
In short
AI search engine tracking identifies the visitors that ChatGPT, Perplexity, Gemini, Claude, Copilot, Grok and DeepSeek send to a website, reports each engine as its own traffic source instead of Direct or Referral, and follows those visitors to the lead and the closed deal in the CRM, so AI search reports in leads and revenue next to paid and organic.
Recognise every engine on arrival
Name the engine behind the visit the moment it lands, from the referrer and the parameters the engine sends, with no channel groups or referrer lists to maintain. See the ChatGPT integration for how one engine is handled.
Record the visit first-party
Read the referrer on your own domain, on the server, before an ad blocker or an iOS setting drops it, so the engine's numbers are not thinned by the browser.
Follow the person to the deal
Keep the AI visit in one journey per person, so a return through Google a week later and the form after it still belong to the engine that started it. See customer journey tracking.
Report revenue per engine
Join the journey to the lead, the pipeline and the closed amount in the CRM, so each engine reports in the same columns as Meta, Google and organic search.
Why your analytics cannot see it
Your prospects say "I found you on ChatGPT". Your analytics says Direct.
The recommendation happens inside the engine, where no tool can look. What arrives is a visit your analytics cannot classify, so the fastest growing channel in B2B reports as nothing.
- GA4's default channels were defined before AI search existed, so an AI referral has no channel to land in and reads as Direct, Referral or Unassigned.
- Clicks out of the ChatGPT app often arrive without a referrer, and a visit without a referrer is direct by definition.
- Gemini sits inside Google, so the clicks it sends blend into your organic search column and inflate your SEO numbers.
- Ad blockers and iOS stop the browser script before the referrer is read, so even the clicks that carry one go uncounted.
- The session ends at the form. Analytics never learns whether the visitor ChatGPT sent became a customer or a closed-lost deal.
- The lead comes back through Google a week later, so last click hands the deal to brand search and AI search gets nothing.
So the channel your prospects keep naming on sales calls is invisible in every budget meeting. Nobody funds a channel that reports zero.
Watch it
Four AI engines, four channels
ChatGPT, Perplexity, Gemini and Claude pulled out of direct traffic and followed through to meetings and revenue.
44 secondsAll videos
Case studies
What changes once the attribution is right
One measured result per customer, with the person, the role and the company behind it.
SEO agency
€500k MRR reached profitably, scaling an SEO agency on performance marketing
Nikita Yatsun
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
LinkedIn growth agency
2x customer retention, once every LinkedIn touchpoint was tracked to revenue
Marc Richard
CEO
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 story
Getreach
Backlinks marketplace (SaaS)
How the data moves
Seven engines in, one journey per person, revenue credited back to the engine.
Each engine sends visits with its own signals. LeadJourney reads them on the server and assigns the visit to its engine. The credit stays there when the conversion arrives through Google a week later.
AI search engines
LeadJourney
AI visits recognised on arrival
What it becomes in your CRM
- Leads
- Booked calls
- Pipeline
- Closed revenue
Engine written onto the record in
and 100+ moreClosed deals credited back to the engine that started the journey
Sources and CRMs
Every AI engine on one side, your CRM on the other.
The engines are recognised without a connection on their side. The CRM is where the leads and the revenue behind each engine come from.
Ad platforms
Spend, campaign structure and click IDs come in. Closed deals go back out as offline conversions.
Meta Ads
Google Ads
LinkedIn Ads
Microsoft Bing Ads
TikTok AdsSoon
Pinterest AdsSoon
Reddit AdsSoon
X AdsSoon- ChatGPT AdsSoon
AI search engines
The traffic your analytics still files under direct. No setup, the tracking script separates it.
CRM and sales
Lead stage and deal value flow in. Every campaign sees the revenue it actually produced.
- HubSpot
- Salesforce
- Pipedrive
- Attio
- Close
GoHighLevel
ActiveCampaign
Odoo
Patch Retention
Podium
monday.comSoon
What the software does
Everything AI search engine tracking has to do, on by default.
Seven engines as their own sources
ChatGPT, Perplexity, Gemini, Claude, Copilot, Grok and DeepSeek each report as a source of their own, never as Direct or Referral, and new engines are added as they appear. Gemini is split out from Google organic.
First-party, server-side recognition
The referrer and the parameters are read on your own subdomain, on the server, which keeps 95%+ of AI referrals visible when a blocker or a consent banner stops the browser script.
Prompt and query context, where the engine passes it
Where an engine sends the prompt alongside the click, it is kept on the visit and shown on the journey, so you can read what the buyer asked. Not every engine passes it.
One journey from the answer to the deal
The AI visit, the return through brand search, the form, the call and the CRM stages sit on one timeline per person, so the engine that opened the journey is visible when the deal closes weeks later.
The engine written onto the CRM record
The source lands on the contact and the deal in HubSpot, Pipedrive and Salesforce, so sales knows a lead came out of a ChatGPT answer before the first call, and your CRM reports name the engine too.
Leads and revenue per engine, next to paid and organic
Each engine reports visits, leads, pipeline and closed revenue in the same table as Meta, Google and organic search. That is the number the GEO and AI SEO work needs.
Why the model matters here
AI search is usually the first touch. Last click gives its deals to Google.
Most AI search leads do not convert on the spot: someone asks ChatGPT, reads your comparison page, then searches your brand a week later and books. Last click hands that deal to brand search. LeadJourney keeps the whole journey, so the engine that started it keeps the credit.
The reports
What you see once every engine has a name.
Your workspace
Lead journeyOne lead, ChatGPT answer to closed won
MSMira SchulzNorthwind LogisticsClosed won
- ChatGPTAsked "best attribution tool for a CRM", clicked the cited comparison pageDay 0
GoogleSearched the brand, read the pricing pageDay 6- DemoBooked a demo from /pricing, contact created in HubSpotDay 9
- HubSpotDeal moved to Closed won, €91,000Day 41
Credited to ChatGPT as first click€91,000
GA4 filed day 0 under Direct and gave the deal to brand search. The journey gives it to the answer that started it.
Lead journey
One lead, from the ChatGPT answer to the closed deal.
Open any lead and read the whole path: the engine that sent the first visit, the prompt where it passed one, the return through Google, the form and the closed amount.
- First visit with the engine, the landing page and the prompt where it was passed
- Return visits across channels and devices in the same journey
- CRM stages and the closed amount on the same timeline
- First click, last click and multi-touch views of the same path
Your workspace
AI search enginesEngine, prompt context and landing page, last 30 days
4,400AI visits
112Leads
€305.0kClosed revenue
EnginePrompt or queryLanding pageLeads
ChatGPT"best attribution tool for a CRM"/comparison/ruler-analytics14ChatGPT"track chatgpt traffic in hubspot"/integrations/hubspot9Perplexity"server-side tracking tools EU"/security8GeminiNot passed/pricing6ClaudeNot passed/blog/server-side-vs-browser-tracking4Where an engine passes nothing, the row still carries the engine and the page. No prompt is inferred.
Engine report
Which engine, which prompt, which page, how many leads.
Every engine in one table with the prompt or query where the engine passed it, the landing page it sent the visitor to and the leads that visit produced. The rows where the prompt is missing say so, rather than modelling one.
- Visits, leads, pipeline and revenue per engine
- Prompt or query context where the engine passes it
- Landing pages per engine, so you see which content gets cited into leads
- Filter any engine to the named leads behind the figure
Your workspace
AI referralsRecognised on arrival, last hour
Live- Visit · /comparison/ruler-analyticsRecognised as ChatGPTNew journeyNamed
- Form · /request-demoRecognised as Gemini, split from Google organicLead created · HubSpotMatched
- Visit · /securityRecognised as PerplexityReturning · day 5Same journey
- Visit · /blog/server-side-vs-browser-trackingRecognised as ClaudeNew journeyNamed
Every row would be "(direct) / (none)" in the analytics tab next to it.
Live feed
AI referrals, recognised as they arrive.
A live log of every AI visit as it lands: the engine, the page, new or returning, and the lead it became. Atlas answers questions about it in Slack or through the MCP server.
- Every AI visit with its engine and landing page, live
- New journey or returning person, stitched at form submit
- Lead created and matched to the CRM, in the same row
- Read-only logs for the whole team, nothing to trust blindly
Live in 21 minutes
AI search tracking set up in three steps
No channel groups to define, no referrer lists to maintain, nothing to tag.
- 1About 5 minutes
Install
Add one script to your site on your own tracking subdomain. From the first page view it recognises AI referrals on arrival and the engines separate from Direct immediately.
- 2About 15 minutes
Connect your CRM
Sign in over OAuth for HubSpot, Pipedrive, Close or Attio, or connect the webhook for Salesforce and the rest, so pipeline and closed revenue join each engine. Add your ad accounts and AI search sits next to paid in the same report.
- 3Same day
Read the engines
From the first visit onwards, ChatGPT, Perplexity, Gemini and Claude report as their own sources, every converting visit becomes a named lead, and the engine is written onto the CRM record.
How it compares
GA4, a hand-built referrer list, and tracking built for AI search.
| What you need |  GA4 default channels | Manual referrer rules | LeadJourney |
|---|---|---|---|
| Names each AI engine as its own source | No: Direct, Referral or Unassigned | Partly: Only the referrers you listed, by hand | Yes: Seven engines recognised on arrival, no setup |
| Catches app clicks that arrive without a referrer | No: No referrer means Direct | No: Nothing for the rule to match | Partly: Recognised where the engine passes a parameter |
| Survives ad blockers, iOS and consent banners | No: Browser script, blocked with the rest | No: Same script, same losses | Yes: Recorded first-party on the server |
| Follows the visit to the lead and the deal | No: Sessions and events | No: Sessions and events | Yes: Named lead, CRM stage, closed amount |
| Reports revenue per engine | No: Not without custom imports | Partly: Sessions per referrer, no revenue | Yes: Leads and revenue per engine, from the CRM |
| Keeps the AI credit when the lead returns via Google | Partly: Data-driven or last click, per session | No: Last session wins | Yes: First click and multi-touch, per person |
What changes
What changes once AI search reports like paid.
AI search is a channel, not a rumour
"Is anyone actually coming from ChatGPT?" gets an answer with visits, leads and closed revenue on it, per engine.
"Direct" traffic
AI visits that arrived without a usable referrer stop landing in the Direct bucket. Each one is assigned to the engine that sent it.
Leads and revenue per engine
ChatGPT, Perplexity, Gemini and Claude are compared to Meta, Google and organic in the same numbers, so the channel defends its own budget.
Sales knows before the call
The engine is on the CRM record, so a rep opening a lead knows it came out of a ChatGPT answer and which page it read.
First touch credit kept
When the lead converts through brand search a week later, AI search keeps the credit it earned as the touch that started the deal.
Content that gets cited gets a number
The pages the engines send buyers to are visible per engine, so the next guide or comparison is a decision, not a guess. Agencies show it to clients as a line item; see SEO agencies.
Original reviews
What our customers wrote, word for word
4.9 out of 5 across 11 public reviews. Quoted as they were left, shortened only by dropping whole sentences.
- 5 out of 5 starsGoogle
“Before LeadJourney, we had no reliable tracking concept for our five-figure ad spend. We were manually building spreadsheet and CRM reports, inaccurate and time-consuming. Within two days, everything was set up. For the first time, I know exactly what I pay per lead and which campaigns actually bring in the best-qualified prospects.”
Florian BuckCEO, Klickkraft GmbH “With LeadJourney we are able to track all our leads and connect them with sales and attribution data to make better decisions. In the first month of using it we scaled from 0 to 100k revenue from paid ads only.”
GetreachBacklinks Marketplace (SaaS)- 5 out of 5 starsG2
“LeadJourney finally fixed my Marketing Analytics. It goes way beyond basic Ad Tracking Software. The Customer Journey Report saves hours of digging, and capturing everything from Offline Conversions to AI Search Tracking makes it the Best B2B Attribution Platform available.”
Sascha LenzMarketing Manager
- 5 out of 5 starsTrustpilot
“Connected LeadJourney for 2 clients, setup took literally 20 minutes each. The data became more accurate, the reports actually make sense. Now clients look at the dashboard and the 'why don't the numbers match?' questions are gone.”
Alexander SamarPerformance Marketing Agency - 5 out of 5 starsGoogle
“With LeadJourney, we have finally found a tool that provides us with the data we need to scale our performance marketing campaigns. The most important KPI is no longer lead price but cost per qualified lead.”
Steffen SiesingCEO, Bilanzmanufaktur GmbH
- 5 out of 5 starsG2
“The ability to track both online and offline conversions in one unified dashboard has given us insights we never had before. Our ROI has improved dramatically since we started integrating LeadJourney with our CRM. We're finally able to see the full customer journey, and it's been a game changer for our strategy.”
Andre WitzelFounder, Trading.de - 5 out of 5 starsTrustpilot
“The ability to seamlessly integrate data from multiple channels and see real-time insights has significantly improved our campaign results. We now focus on metrics that truly matter, like ROI and qualified leads.”
Nikita YatsunCEO, RLV Media GmbH
Read the reviews where they were left
Common questions
AI search engine tracking, the questions buyers ask
What marketing teams want to know before they put a number on ChatGPT.
Why does ChatGPT traffic show up as Direct in GA4?
GA4 classifies sessions with channel definitions written before AI search existed, so an AI referral has no channel of its own to land in. A click from chatgpt.com in a browser is a Referral, a link with a source parameter and no medium is Unassigned, and a click out of the ChatGPT app often arrives with no referrer at all, which is Direct by definition. Nothing is broken; GA4 simply has no ChatGPT bucket.
Which AI engines does LeadJourney track?
ChatGPT, Perplexity, Gemini, Claude, Copilot, Grok and DeepSeek are reported as their own sources instead of landing in Direct or Referral. Gemini is split out from Google organic. New engines are added as they appear, with nothing for you to change.
Do I need to configure anything to track AI search traffic?
No. AI search tracking is on by default. The same script that records your sessions recognises AI referrals on arrival, server-side, so there are no channel groups to define, no referrer lists to maintain and nothing to tag. Most teams are live in about 21 minutes, CRM included.
Can I see the prompt or the question someone asked?
Where the engine passes it, yes. Some engines send the prompt or query context alongside the click, and LeadJourney keeps it on the visit and shows it on the journey and in the engine report. Not every engine passes it, and where one does not, the visit still carries the engine and the landing page. We show the rows where the prompt is missing as missing rather than modelling one.
Can you track Google AI Overviews too?
Honestly, no one can reliably. A click from an AI Overview arrives with a normal google.com referrer, so it is technically indistinguishable from a regular organic Google click. A tool that claims to split out AI Overviews is guessing. LeadJourney tracks the engines that send identifiable signals, and AI Overview clicks stay part of your Google organic traffic, tracked to the lead and the deal like every other visit.
Can I see revenue from ChatGPT, not only visits?
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 the pipeline and the closed revenue in your CRM. So each engine reports visits, leads, pipeline and won revenue, the same columns as your paid channels, and the engine is written onto the CRM record.
What about a lead who finds us in ChatGPT and comes back through Google a week later?
That is the case most tools get wrong. Last click credits the branded Google search and AI search reports nothing. LeadJourney keeps the whole journey per person, so the ChatGPT visit and the Google return are one path, and you can read it under first click, last click, linear, position-based or multi-touch. Under first click, the engine that started the deal keeps the credit.
How is this different from AI visibility tools that track citations?
Citation checkers tell you which prompts mention you and how often. That is the top of the funnel, and it is useful. AI search engine tracking measures what happens after the click: the visit, the lead, the pipeline and the revenue per engine, from your CRM. Most teams run both; the ranked guide compares the tools in each group side by side.
See it on your own traffic
See how many leads AI search already sends you.
Book a demo and we split your own Direct traffic into ChatGPT, Perplexity, Gemini and Claude live, with the leads behind each engine.
Your workspace
Where AI search traffic was hidingSame 30 days, same visits, two reports
What GA4 could name880visits with a usable referrer; 3,520 more sat in Direct and Unassigned
What LeadJourney named4,400visits assigned to the engine that sent them, 112 of them leads
Recovered from Direct+3,520 visits named
EngineGA4LeadJourney
ChatGPT2101,840Perplexity4801,120Gemini90640Copilot60412Claude40388Recognised server-side from the signals each engine sends, not guessed from a referrer that is usually missing.
