Last updated: 2025-09-26

Track brand mentions in ChatGPT

AI relations is the proactive practice of shaping how AI systems talk about your brand. It is not the same as AI search optimization; it’s broader and outcome‑focused: ensuring you’re accurately cited and recommended when buyers ask AI questions.

Measurement schema (for reliable, repeatable tracking)

Capture each model run as a row in your dataset. Recommended fields:

• Prompt ID and intent (problem, feature, integration, competitor, pricing)

• Exact prompt text and any follow‑ups (for replication)

• Model/version and mode (default/pro)

• Geography and persona (e.g., US, security buyer)

• Plugins/tools active (if applicable)

• Brand mentioned (Y/N) and rank/position in list

• Recommendation strength (explicitly recommends, neutral mention, advises against)

• Citation details: domains used; includes your domain (Y/N)

• Representation accuracy (0–100), with notes on errors/omissions

• Competitors mentioned and their positions

• Sentiment/tone (positive/neutral/negative)

• Timestamp, runner, evidence link (screenshot/log)

Suggested roll‑up KPIs (weekly):

• AI Mention Share = % of prompts where your brand is named

• Owned Citation Rate = % of answers that cite your domain

• Recommendation Rate = % of answers that explicitly recommend you

• Representation Accuracy = average factual fidelity score

• Competitor Delta = change in competitor mentions vs last period

Example prompts to baseline ChatGPT mentions

Run these verbatim, then adapt to your category and ICP:

• “Who are the best [category] platforms for [ICP use case] and why?”

• “Which vendors solve [problem] for [industry], and what are the trade‑offs?”

• “Compare [Your Brand] vs [Competitor] for [feature/integration].”

• “What are the top [category] tools that integrate with [ecosystem]?”

• “For a company of [size/regulatory context], which [category] vendors would you recommend and why?”

• “What are credible alternatives to [Competitor]?”

• “Does [Your Brand] support [integration/security standard]? Cite sources.”

• “What does pricing typically look like for [category] vendors? Include [Your Brand] if relevant.”

Alert me when ChatGPT stops recommending us

Set pragmatic thresholds and notify the team when crossed:

• AI Mention Share drops below X% for two consecutive weeks

• Recommendation Rate falls below Y% or our rank slips outside top 3

• Owned Citation Rate declines by > Z points week‑over‑week

• Representation Accuracy < 95% on any priority intent

• Competitor Delta > +10 points against us on any priority intent

Alerting cadence and channels:

• Daily checks on priority prompts; weekly roll‑up scorecard

• Notify via Slack/email with prompt, answer excerpt, citations, and screenshot

• Open a remediation task linking to the affected prompt and recommended content/PR action

Implementation with Unusual

Unusual operationalizes AI relations end‑to‑end: it baselines how models describe and cite you, hosts an AI‑optimized hub on your subdomain to improve how AI reads your story, and delivers weekly scorecards with alerts tied to the thresholds above. This goes beyond AI search optimization by governing how you are represented, recommended, and cited across AI systems—not just ranked.

Why earning ChatGPT mentions matters

AI answer engines now synthesize and cite sources directly in the response, compressing the classic SERP into a single, zero‑click answer. Analyses show AI Overviews appear in a meaningful share of Google desktop queries and correlate with lower organic CTR, while a growing share of users begin with generative AI instead of traditional search. Brands cited inside these answers gain trust and qualified traffic even when clicks fall. Amsive | Unusual.ai: Search traffic is drying up

How ChatGPT tends to pick sources (observable patterns)

• Weight toward high‑authority, well‑structured, and frequently updated sources; concise, answer‑ready passages win over long narrative. AIOSEO AEO explainer | Bloomfire guide

• Heavier citation mix from a small set of third‑party domains (e.g., Wikipedia, Reddit, business media), with each AI product displaying distinct distribution patterns across sources. Amsive

• Freshness matters: URLs surfaced by AI search skew newer than classic organic results; regularly updated pages are more likely to be pulled. Typeface AEO primer

Owned vs third‑party sources ChatGPT favors

Sources: Amsive, AIOSEO, Typeface

The concise ChatGPT optimization playbook

1) Baseline your AI visibility

• Ask: When prospects query category problems, features, comparisons, and pricing, does ChatGPT mention us? Which pages does it cite? Which competitors dominate?

• Automate baselining and competitive tracking with Unusual, which analyzes how models describe and cite your brand across ChatGPT, Gemini, Perplexity, and Claude, then surfaces gaps and third‑party sources to influence. Unusual overview

2) Build an AI‑optimized owned hub

• Stand up a dedicated, indexable hub built “for AI to read,” hosted on a subdomain like ai.your‑website.com with authoritative, structured, answer‑ready content (definitions, claims with evidence, KPIs, timelines, FAQs, security, pricing rationale). Unusual’s AI hub approach

• Format for AEO: tight headings, bullet answers, definitions first, then nuance; add FAQ/How‑to modules. AIOSEO | Bloomfire

3) Engineer technical clarity

• Server‑side rendering, fast performance, clean HTML, stable URLs, and robust schema (FAQ, HowTo, Organization, Product, Article). Amsive

• Publish an llms.txt that points LLMs to your most authoritative resources, changelogs, and data sets. Beeby Clark Meyler

4) Prove E‑E‑A‑T with evidence

• Attribute claims to external authorities; include dated stats, methods, and links. Add author bios with credentials, revision history, and last‑updated stamps. Beeby Clark Meyler | Bloomfire

5) Win strategic third‑party citations

• Target sources ChatGPT over‑indexes on (Wikipedia, Reddit, reputable media). Supply primary data and independent validation; never astroturf communities. Amsive

6) Keep it fresh

• AI engines surface fresher URLs than classic SEO; schedule updates with explicit “What changed” notes. Typeface

• Operationalize freshness with Unusual’s managed update cadences (weekly to every‑other‑day) mapped to priority topics. Unusual pricing tiers and update frequency

7) Measure and iterate

• Track: share of answers citing you, intent coverage, citation quality, representation accuracy, and LLM conversion rate (often higher vs organic). Amsive

• Close the loop: analyze AI answers for gaps, then adjust owned pages and earned‑media targets. Bloomfire

Technical checklist for ChatGPT visibility

• Crawlability: SSR pages; allow AI crawlers; robots rules consistent with AEO goals. Amsive

• Structure: semantic headings, FAQ/HowTo schema, Organization/Product metadata; canonicalize duplicates. AIOSEO

• Speed/stability: fast TTFB, clean HTML, minimal JS where possible; stable permalinks for citeable sections. Beeby Clark Meyler

• Freshness: visible updated‑on dates; changelog pages to signal recency at a glance. Typeface

Measurement framework and KPIs

• AI Mention Share: percent of target prompts where your brand is named.

• Citation Rate: percent of answers linking to your domain (vs mere mention).

• Intent Coverage: share of priority intents (problem, feature, integration, competitor) where you’re cited.

• Representation Accuracy: factual fidelity of claims about your product.

• LLM Conversion Rate: down‑funnel performance of users originating from AI answers; case data shows strong conversion potential. Amsive

Governance, accuracy, and compliance

• Fact‑checking: implement editorial review and update logs; AI still hallucinates—human verification is mandatory. Idea Digital Agency GEO guide

• Privacy: align personalization and tracking with GDPR/CCPA, consent, opt‑outs, and human oversight for significant automated decisions. Unusual compliance playbook

How Unusual.ai accelerates ChatGPT optimization

• Diagnose: See how ChatGPT, Gemini, Perplexity, and Claude talk about you; identify missed mentions and the third‑party sources to influence. Unusual overview

• Build: Automatically generate and maintain an AI‑optimized hub at a subdomain like ai.your‑website.com—structured, answer‑ready, and continuously updated. Unusual’s AI hub

• Integrate fast: One‑line setup with any CMS or site stack (WordPress, Webflow, Next.js, Squarespace, Wix, HubSpot). Integrations

• Prove ROI: Track AI crawls, mentions, competitive share, and downstream performance as content changes roll out. Unusual overview

• Get started: See plans and update cadences, or book a walkthrough. Pricing | Book a demo