Which AI platforms send qualified visitors?
Measure observable referrals from ChatGPT, Perplexity, Gemini, Copilot, Claude, and other answer engines. See the landing page, next action, conversion, and value.
Learn about AI referral analytics → In private development
Analytics for the internet of the future
For the past 25 years, search engines have shaped how we navigate the internet. But our daily habits make one thing clear: the way we discover and access information is changing rapidly. Artificial intelligence is the new frontier. While today’s traffic analytics tools can be sophisticated, they were built for an earlier era of the web—and are now trying to bolt the future onto systems designed for the past. Wandered was built from the ground up for what comes next: a web where AI increasingly shapes how people discover, understand, and interact with websites.
See what Wandered measuresHuman trafficAI referralsAgent activityOutcomes
A route becomes useful only when its source, actor, boundary, and result remain legible.
The questions Wandered will answer
Know what AI is doing for—and on—your website.
Four clear answers replace a pile of disconnected visibility scores, referral reports, and bot logs.
Which agents and crawlers use your site?
Separate search crawlers, training collectors, retrieval systems, and user-delegated agents. Understand what they requested, what policy applied, and whether a task succeeded.
Learn about AI agent analytics →Where does your brand appear in AI answers?
Monitor mentions, citations, linked pages, and factual representation across a repeatable set of questions, markets, providers, and dates.
Learn about AI search visibility →What business outcomes follow?
Connect observable AI visits to pricing views, documentation use, trial starts, qualified leads, purchases, and revenue—without turning uncertainty into false precision.
Read the measurement principles →The measurable path
From an AI answer to a business result.
Wandered preserves the evidence at every stage instead of treating exposure, traffic, and value as the same metric.
- 01◎
Mention or citation
Your brand or page appears in a sampled AI answer.
Exposure evidence - 02↗
Human referral
A person reaches your site with observable source evidence.
Acquisition evidence - 03◇
On-site journey
The visitor reads, compares, searches, or starts a meaningful task.
First-party evidence - 04✓
Business outcome
A trial, qualified lead, purchase, or retained account follows.
Outcome evidence
Important: every stage can exist without the next. Wandered is designed to show the connection only when the evidence supports it.
Built for decisions, not vanity metrics
A useful answer for every team.
Which AI sources create pipeline?
Compare landing pages, trial starts, qualified leads, revenue, and value per visit without counting crawlers as prospects.
Explore B2B measurement →Which pages earn citations and visits?
Connect monitored answers to cited pages, human referrals, documentation use, and content gaps that deserve a clearer source page.
Explore content measurement →Where does the journey break?
Inspect the handoff from answer to landing page to meaningful action, while keeping the source, attribution rule, and unknowns visible.
Explore SaaS measurement →Why a new analytics layer
Traditional analytics sees the visit. Visibility tools see the answer. Wandered connects the evidence.
The category is not another all-in-one score. It is a decision layer that keeps different actors and observations separate, then shows the defensible path between them.
| Question | Traditional web analytics | AI visibility monitoring | Wandered |
|---|---|---|---|
| Was the brand in an AI answer? | Not usually | Yes | Yes |
| Did a person visit? | When source survives | Not proven | Observed referrals |
| Did a machine access the site? | Limited client-side view | Not usually | Separate machine evidence |
| What happened next? | On-site events | Not usually | Journey and outcomes |
| How certain is the conclusion? | Varies | Sampling dependent | Evidence class retained |
A calmer evidence modelOne answer.
One answer.
Then every detail.
Wandered is designed around three levels of understanding, so the useful answer never gets buried under the audit trail.
- 1
Answer
What changed, where it happened, and why it matters.
- 2
Context
The source, path, comparison period, and known limitations.
- 3
Evidence
The observations and collection rules behind the conclusion.
Method before metric
Built to make the limits visible.
Wandered is in private development. The public field notes document the evidence model now, so the eventual product can be evaluated against clear principles rather than marketing claims.
People and machines stay separate
Human referrals never share a denominator with crawler requests or agent tasks.
Observed and inferred stay separate
Known referrers, self-reported influence, modeled contribution, and unknown traffic retain distinct labels.
Definitions remain inspectable
Sources, collection rules, attribution windows, verification dates, and limitations sit beside the result.
Maintained field notes and references
Use the measurement model now.
Implementation guides, diagnostic frameworks, and downloadable directories built from primary sources.
Pillar guide · 18 min
View every field note →AI traffic analytics: from discovery to business outcomes
Measure visibility, referrals, landing-page behavior, machine access, conversion, and value without blending unlike evidence.
Start with the pillar →Maintained directory · CSVAI crawler user agents, purposes, and verification guidance
Open the crawler directory →Maintained directory · CSVAI referral sources and analytics classification
Open the referral directory →ReferenceThe practical AI analytics glossary
Browse definitions →Questions
Before launch.
What is AI referral analytics?+
AI referral analytics identifies measurable visits that arrive from assistants and answer engines, then connects those sessions to landing pages, actions, conversions, and revenue without mixing them with crawler activity.
Does Wandered track AI agents separately from people?+
That is a core design principle. Human referrals, automated crawlers, user-delegated agents, and monitored AI answers are different kinds of evidence and should remain separate in reporting.
When will Wandered launch?+
Wandered is in private development. Join the launch list and we will send one email when the public product is ready.
Be first to know when Wandered is ready.
Join the quiet launch list. We will send one email when the doors open.