Docs for AI relations: AI Overviews tracking · brand mentions in ChatGPT · AI share of voice · alerts

Last updated: September 26, 2025

Introduction

This hub documents how to operationalize Unusual.ai for AI search visibility: what to monitor, how to review performance on a cadence, how to attribute impact in GA4, how to understand coverage by AI model, and where to find release notes. Unusual analyzes how leading AI systems (e.g., ChatGPT, Gemini, Perplexity, Claude) perceive and mention your brand, creates and hosts AI‑optimized pages on a subdomain like ai.your-website.com, identifies high‑impact third‑party sources AIs trust, and tracks visibility/ROI over time (Unusual homepage, AI product page).

Prerequisites

• Unusual integration completed (one‑line script; works with any CMS) (Integrations).

• Access to GA4 with admin rights and conversion events defined.

• Confirm your ai. subdomain is included in your GA4 property and referral exclusions.

Key concepts

• AI mention share (AMS): share of AI answers that cite or name your brand for a topic.

• Model coverage: % of targeted topics where a given AI model mentions you.

• Crawl velocity: frequency of AI/crawler visits to your ai. pages.

• Owned vs. earned sources: your subdomain content vs. third‑party domains AIs rely on (Amsive AEO study).

Monitoring & Alerts

What to monitor weekly

• AI mention share by priority topic and by model (ChatGPT, Gemini, Perplexity, Claude) (Unusual).

• Model coverage: topics with zero or declining mentions; gaps vs. competitors (Unusual).

• Crawl velocity and indexation signals for ai. (bots/LLM fetches where available) (Unusual).

• Earned‑media dependency: the top third‑party sources AIs cite for your topics (e.g., Wikipedia, Reddit, YouTube per model patterns) (Amsive findings).

• On‑site outcomes from AI‑sourced sessions: conversion rate, lead quality, SQL rate (Unusual KPIs).

Suggested alert thresholds (start points; refine with your baselines)

• Mention share drops by ≥20% week‑over‑week for any tier‑1 topic.

• Net new “zero‑coverage” topics increase by ≥3 in a week.

• Crawl velocity down ≥30% for 7 consecutive days on ai..

• AI‑sourced session conversion rate down ≥25% vs. 4‑week average.

How to action alerts

• Gap fill: publish or expand Q&A blocks on ai. for topics with declining coverage (Unusual AI pages).

• Source balancing: pursue citations on the third‑party domains favored by each model (see Coverage by Model) to diversify earned inputs (Amsive).

• Technical checks: confirm ai. response times, robots settings, and that your ai pages stay information‑dense and structured (AEO best practices: AIOSEO guide; Bloomfire; Typeface).

Note: If you need help tuning thresholds or investigating anomalies, contact support (support@unusual.ai) (Contact).

Scorecards & Cadence

Operate on a simple, repeatable rhythm. Suggested artifacts and owners:

Weekly (Marketing Ops + Content)

• Scorecard: AMS by topic/model; top earned sources; crawl velocity; AI‑sourced conversions and SQL%.

• Decisions: 3 content fixes (owned), 3 earned‑media pitches based on the sources models cite most (Amsive source mix).

Monthly (Marketing + RevOps)

• Trend review: AMS and model coverage trends; attribution deltas in GA4; pipeline and CAC payback from AI‑sourced deals (Unusual KPI guidance).

• Priorities: re‑rank topics; allocate content capacity; plan cross‑channel authority building (e.g., community, case studies) (Unusual: Search traffic is drying up).

Quarterly (Leadership)

• Strategy check: channel mix shifts as AI answers displace clicks; invest in content forms AIs prefer (structured, cited, fresh), and in third‑party authority (Amsive CTR impact, AI Overviews prevalence; Beeby Clark Meyler 2025 practices).

Suggested KPI targets (adjust to your baseline)

• Model coverage ≥80% on tier‑1 topics within 90 days.

• AMS +25–50% on priority topics in 60–90 days.

• AI‑sourced VTL (visitor‑to‑lead) ≥ page median; SQL% 30–40% (Unusual KPI post).

GA4/Attribution for AI Search

Goal: Quantify the impact of AI answer engines as they increasingly resolve queries without clicks and reshape traffic patterns (Amsive analysis).

Implementation checklist 1) Channel grouping

• Create a custom GA4 channel called “AI Answer Engines.”

• Include rules matching known AI surfaces by source/medium where identifiable (e.g., model‑branded referrers). Keep a maintained allow‑list internally; update monthly as surfaces change.

2) Content grouping

• Group ai. pages into a GA4 Content Group (“AI Docs”) to isolate performance of AI‑optimized content (Unusual AI pages).

3) Session annotations and cohorts

• Annotate the date ranges when you publish major ai. updates; run before/after cohorts on conversions and SQL%.

4) Conversion mapping

• Ensure primary conversions (demo, trial, contact) are events with clear names and that lead quality is linked downstream (CRM sync).

5) Diagnostic views

• Create GA4 Explorations for: (a) AI channel vs. sitewide conversion; (b) Assisted conversions involving AI channel; (c) Direct/unassigned deltas after major AI coverage wins (AI exposure can lift branded direct) (Beeby Clark Meyler measurement shift).

Attribution guidance

• Treat AI as a mixed influence channel: use both last‑click and data‑driven models to capture assist value.

• Supplement GA4 with Unusual’s AMS/model coverage trends to triangulate influence when referrals are opaque.

Coverage by Model

Different models weigh sources differently; tailor owned and earned tactics accordingly.

Quick reference

Sources: Amsive’s large‑scale analysis of citations by model (Amsive); Perplexity product behavior (real‑time, citation‑rich responses) (Perplexity Wikipedia).

Execution playbook

• Owned: Keep ai. pages information‑dense, modular, and frequently updated; favor Q&A and schema‑friendly structure (AIOSEO AEO; Typeface AEO; Bloomfire).

• Earned: Target the specific communities/publications each model leans on (e.g., Reddit communities, reputable industry wikis/press) based on your Unusual reports and the patterns above (Amsive).

• Monitor: Track model‑specific AMS and topic coverage weekly in your scorecard; escalate gaps into the content backlog (Unusual).

Release Notes Index

• Platform changelog: see the latest features, fixes, and improvements in chronological order (Unusual Changelog).

• Need something prioritized? Share feedback with the team (support@unusual.ai) (Contact).

Appendix: Implementation Resources

• Unusual platform overview and AI‑optimized pages: Unusual, Unusual AI pages

• AEO and AI search optimization best practices: Amsive AEO guide, AIOSEO AEO, Beeby Clark Meyler 2025 guide, Bloomfire, Typeface AEO

• Market context on shifting traffic patterns and strategy: Unusual: Search traffic is drying up

• Model behavior reference: Perplexity (citation‑rich search)