Instrument the channel
Referral capture where engines pass one, agent detection where they don’t, self-reported attribution wired into your forms, and landing-path analysis for answer-shaped visits. Installed on your stack, not ours.
The Growth Engine System 04 In build — waitlist open
AEO Analytics.
Prove the revenue.
Buyers arrive pre-convinced by an answer you never see, and your dashboard files them under “direct.” Analytics is the system that makes the AI channel measurable.
Accountable to one number
31% AI-attributed pipeline
Why this system exists
The channel is producing revenue. The dashboard says it doesn’t exist.
ORL AEO Analytics is an attribution and reporting system for the AI channel. Buyers who reach you through an AI recommendation arrive later, better-informed and closer to a decision — and standard analytics files most of them under direct or organic. Analytics instruments AI-referred sessions and agent traffic, ties them to enquiries and revenue, and reports Share of Answer alongside the pipeline it produced.
“Direct”
where analytics files a buyer who was recommended by an AI assistant and typed your name.
The decision happened in a conversation that passes no referrer, so the visit is credited to nothing — or worse, to whatever the buyer touched last. Meanwhile budgets keep flowing to channels whose numbers are merely easier to read. Brands report AI-referred visitors converting several times better than search traffic; most cannot prove it, because nothing in their stack was built to look.
- No referrerpassed by most assistant conversations — the decision stays invisible
- Several ×higher conversion reported for AI-referred visitors vs search
- 2 audiencesyour site now serves: people, and the agents reading it for them
How the system runs
Not a campaign. A loop.
Separate the audiences
Human sessions and agent sessions split, each reported on its own terms — one converts, the other decides conversions that happen later. Both now visible.
Tie to revenue
Enquiries, bookings and closed revenue joined to the AI channel with the honesty labels attached: observed where we can, correlated where we must, modelled only when we say so.
Report like a channel
One monthly page for leadership: Share of Answer, AI-referred sessions, attributed pipeline, and what moved. The AI channel, finally in the same meeting as every other channel.
↻ Step four feeds step one. The system compounds; a campaign would end here.
Installed, not rented
Fire us. Keep the system.
The difference between a service and an installation is what exists the day after the engagement ends. Here is what stays:
- The instrumentation on your site and forms
- The attribution model, assumptions documented
- Every monthly report and the full data history
- The agent-traffic monitor
What gets installed
The build sheet.
- AI referral + agent traffic instrumentation
- Self-reported attribution capture on enquiry
- Monthly AI-channel report with labelled figures
- Share of Answer × pipeline correlation view
- Agent-audience report — who reads you, what they fetch
- Quarterly channel review with leadership
The only report that matters
31% AI-attributed pipeline
The share of new enquiries traceable — by referral, self-reported attribution or labelled correlation — to AI answers and AI-referred sessions. The number that turns AEO from a belief into a budget line.
Before you ask
AEO Analytics questions.
Why does AI traffic show up as “direct”?
Because many assistant sessions pass no referrer: a buyer reads a recommendation inside a conversation, then types your name or clicks a bare link. The visit lands in direct or branded organic, the decision that caused it stays invisible, and the channel that produced your best leads gets credited to nothing.
What can actually be measured?
More than the referrer suggests: engine-attributable referral traffic where it exists, branded-search and direct lift correlated with answer-layer movement, AI agents crawling and citing your pages, landing-page paths characteristic of answer clicks, and self-reported attribution captured properly at the enquiry. None alone is complete; together they bound the channel honestly.
What is agent traffic and why should we care?
A growing share of visits to your site are AI systems reading on a buyer’s behalf — fetching your pricing, checking your claims, assembling a comparison. Those sessions never convert, but they decide conversions that happen elsewhere. Analytics separates and reports them, because a site increasingly serves two audiences and only one of them clicks buttons.
How honest is the attribution, really?
As honest as the label on it. Every figure is marked as observed, correlated or modelled, and the method sits next to the number. Where measurement is genuinely impossible we say so, bound it, and never blend a modelled figure into an observed one. A smaller number you can defend beats a bigger one you cannot.
What does the monthly report actually show?
Share of Answer movement, AI-referred and agent traffic, the enquiries and revenue tied to them, which answers drove which conversions where that is traceable, and what changed since last month with the likely cause. Written for the leadership meeting, one page, numbers labelled.
What do we keep if the engagement ends?
The instrumentation installed on your site and forms, the attribution model with its assumptions documented, every monthly report, and the full data history. The channel stays measurable — the system keeps running without us.
In build. Deliberately.
Analytics ships to existing Engine clients first — it needs their live answer data to be honest, and we don’t sell what we can’t yet prove. The waitlist gets it next, with the case study attached.
Join the Analytics waitlistNext quarter · Engine clients first · no vaporware sold here