GA4 and web analytics
Sessions, pages and traffic sources, including identifiable AI referrers.
Integrations · Analytics and data
Connect your data, without mixing up signals.
Aive can read analytics sources such as GA4 and first-party events to link publishing, AI referrals and conversions. Every signal keeps its level of proof: a crawl is not a citation, a referral is not a cause.
Last verified · 26 September 2026
Which analytics data can Aive connect?
Depending on the sources connected, Aive can link GA4 or first-party data, publishing and Player events, sessions referred by identifiable AI assistants and declared conversions. Each source is to be validated in your workspace. Signals stay separate: crawls, observed citations, referrals and conversions do not say the same thing and are never added together.
- AI referral
- A session arriving from an identifiable AI referrer, observed in your analytics.
- First-party
- Data collected directly on your own properties, with the required consent.
- Assisted conversion
- A conversion with exposure or referral, under a declared attribution model.
Sources
Four families of data.
First-party events
Publishing, Player and engagement events, collected with consent.
AI referrals
Sessions from identifiable AI assistants, kept separate from crawls and citations.
Exports
Data and reports to your BI tools or warehouses, via the Enterprise API.
See the Enterprise API
Levels of proof
Crawl, citation, referral, conversion: four proofs.
| Signal | What it establishes | What it does not establish | Typical source |
|---|---|---|---|
| Crawl | An identifiable bot visited a page. | A citation or a human visit. No log does not prove no visit. | Server logs, tracking |
| Observed citation | An AI answer citing a URL or domain, in a panel and on a date. | A click, a conversion or every user’s experience. | Engine observation by Aive |
| Referral | A visit observed from an identifiable AI referrer. | The prompt or answer that caused it. | GA4, first-party |
| Conversion | A business event linked under a declared model. | On its own, an incremental effect. | CRM, analytics |
Principles from the Aive methodology. No causal statement without an appropriate protocol.
Mechanism
From connected sources to measurement rails.
01Inputs
- GA4 or first-party analytics properties
- Publishing and Player events
- CRM conversion data, if provided
- Analysis period and scope
02Processing
- Source connection and checks
- Separation of crawls, citations, referrals and conversions
- Matching by content, version and period
- Coverage and missing-data checks
03Outputs
- Dashboards per measurement rail
- Observed AI referrals
- Assisted conversions under a declared model
- Exports subject to configuration
Demonstration
Every channel has its own proof.
An illustration of measurement rails: each result keeps its level of proof and is never added to the others.
Social Views, retention, shares and comments per version and platform. Platform signal
AI answers Mentions, citations, sources and contextual sentiment, on a panel and period. Protocol observation
Referral Sessions from identifiable AI referrers, separate from crawls and citations. Observed visit
Conversion Leads, sales and first-party events, under a declared attribution model. Declared attribution
Citation ≠ click · referral ≠ causality · a linked conversion alone does not establish incremental effect.
Control
Measured, separated, dated.
What is measured
- Sessions referred by identifiable AI assistants
- First-party publishing and Player events
- Assisted conversions under a declared model
- Coverage and period of each source
Controls
- Consent for first-party measurement
- Signals kept separate, never added up
- Denominator and period displayed
- Role-based access
Prerequisites and availability
- Access to analytics properties
- Tagging plan or defined events
- Status to be validated by source
- Exports via the Enterprise API on quotation
What this data does not prove
- A referral reveals neither the prompt nor the answer that caused it.
- Unidentifiable AI referrers are not attributed.
- No causal ROI without appropriate instrumentation and protocol.
- Result quality depends on tagging and on the sources connected.
Frequently asked questions
Does Aive replace GA4?
No. Aive reads existing sources such as GA4 or first-party events to link them to content and observed answers. Your analytics tool remains the reference for your sessions.
Does a referral from an AI assistant prove my page was cited?
No. A referral establishes a visit observed from an identifiable referrer, not the prompt or answer that caused it. Citations are measured separately, in a dated panel.
Can I export the data?
Exports to your BI tools or warehouses go through the Enterprise API on quotation, with scopes, limits and support defined in the contract.
Let’s scope your data sources.
The demo checks your stack; sources, tagging and scope are confirmed during scoping.