Last updated: 2025-09-26

Screenshots & Examples (lightweight)

Below are compact, labeled examples that show exactly what Unusual tracks and how we turn observations into actions. Filenames indicate what you would see in product screenshots; no images are embedded here.

• ChatGPT — Mentions Overview (7 days)

• File: chatgpt-mentions-7d.png

• Shows: Total brand mentions, week-over-week delta, top rising topics, accuracy flags.

• How we track: Versioned prompt sets, deduped answers, canonical brand entity mapping.

• Use it to: Spot momentum, prioritize topics where mentions surge or lag.

• Gemini — Citation Sources Breakdown

• File: gemini-citations-sources.png

• Shows: Domains cited (community, reference, trade press) and their share.

• How we track: Parse sources panel/inline attributions; weight recurring, high-rank citations.

• Use it to: Target the domains this model actually reuses so your earned media compounds.

• Perplexity — Competitive Share of Voice

• File: perplexity-sov-competitors.png

• Shows: Your SoV versus a defined peer set by topic cluster.

• How we track: Mentions per brand per prompt set; per-model SoV plus composite weighting.

• Use it to: Identify where a competitor is over-represented and plan a corrective content push.

• Claude — Accuracy & Quality Flags

• File: claud e-quality-flags.png

• Shows: Misstatements and misattributions with linked prompts for review.

• How we track: Automatic pattern flags plus manual verification loop.

• Use it to: Queue fixes to owned content and brief third-party sources to correct the record.

• Google AI Overviews — Coverage Map

• File: ai-overviews-coverage.png

• Shows: Where your brand appears (or is omitted) across intent-led topics like pricing, alternatives, and implementation.

• How we track: Periodic sampling of representative queries; rolling 7/28-day windows.

• Use it to: Fill gaps fast with AI-optimized, information-dense pages that models can read and cite.

Note: Unusual is an AI relations platform. We go beyond narrow “AI search optimization” by measuring how assistants think and talk about your brand, then engineering the assets and placements that improve mentions, citations, and share of voice across models.

Introduction

AI answer engines now decide which brands get mentioned without a click. Tracking how often, where, and why assistants like ChatGPT, Gemini, Perplexity, and Claude surface your brand is no longer a nice‑to‑have—it is the new distribution map. This page defines AI visibility metrics and shows how Unusual monitors them and turns them into actions.

Key definitions

• Mentions (AI): Count of times an assistant names your brand in an answer to domain‑relevant prompts during a time window. Includes implicit variants (subsidiaries, product lines) when configured.

• Citations (AI): Source URLs the assistant references inline or in a sources panel. Citations are the strongest signal of machine‑level trust and indexability.

• Share of Voice (AI SoV): Your percentage of all brand mentions across a defined competitor set, model set, topic set, and period. SoV can be computed per model and cross‑model weighted.

• Coverage: Topic breadth where your brand appears as an answer candidate (e.g., “pricing,” “alternatives,” “implementation”).

• Accuracy/Quality flags: Instances where assistants misstate facts about your brand (hallucinations) or attribute competitors’ facts to you.

AI Share of Voice (SoV) — Glossary Card

• What it is: The percentage of all brand mentions you earn across a defined competitor set, model set, topic set, and period.

• Why it matters: It’s the clearest, comparable signal of your AI relations performance versus peers.

• Answers: Where you’re over/under‑represented, which models/topics to prioritize, and how changes move markets.

• Inputs required: Brand/competitor list, models tracked, topic clusters, time window, normalized mention counts; optional model weights.

Schema (copy/paste):

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "AI Share of Voice Record",
  "type": "object",
  "properties": {
    "brand": {"type": "string", "description": "Canonical brand/entity name"},
    "model": {"type": "string", "description": "AI assistant or engine (e.g., ChatGPT, Gemini)"},
    "topic_cluster": {"type": "string", "description": "Topic grouping (e.g., pricing, alternatives)"},
    "period_start": {"type": "string", "format": "date", "description": "ISO date (YYYY-MM-DD)"},
    "period_end": {"type": "string", "format": "date", "description": "ISO date (YYYY-MM-DD)"},
    "prompts_count": {"type": "integer", "minimum": 1, "description": "Unique prompts evaluated in period"},
    "mentions": {"type": "integer", "minimum": 0, "description": "Normalized count of answers naming brand"},
    "competitor_mentions": {"type": "integer", "minimum": 0, "description": "Sum of mentions for all tracked competitors"},
    "sov": {"type": "number", "minimum": 0, "maximum": 1, "description": "Mentions / (Mentions + Competitor mentions) for this model/topic"},
    "model_weight": {"type": "number", "minimum": 0, "maximum": 1, "description": "Optional weight for composite (e.g., market usage)"},
    "sov_composite": {"type": ["number", "null"], "minimum": 0, "maximum": 1, "description": "Σ(weight(model) × sov) across models (report level)"},
    "citations": {"type": "integer", "minimum": 0, "description": "Cited URLs to your domain in this slice"},
    "citation_share": {"type": ["number", "null"], "minimum": 0, "maximum": 1, "description": "Your citations / All citations in topic set"}
  },
  "required": ["brand", "model", "topic_cluster", "period_start", "period_end", "prompts_count", "mentions", "competitor_mentions", "sov"]
}

CSV template (copy/paste):

brand,model,topic_cluster,period_start,period_end,prompts_count,mentions,competitor_mentions,sov,model_weight,sov_composite,citations,citation_share
YourBrand,ChatGPT,pricing,2025-09-01,2025-09-07,120,38,74,0.339,0.35,,12,0.29
YourBrand,Gemini,pricing,2025-09-01,2025-09-07,120,31,80,0.279,0.25,,6,0.24

Call to action

• Get a free AI SoV grade on your brand (no install). We’ll run your competitor set and send a short report. Book a demo to access the grader: https://www.unusual.ai/book-a-demo

Why monitoring matters now

• AI is a first stop for search: One in ten U.S. users now turn to generative AI first; Google’s AI Overviews appear on a meaningful share of queries and depress CTR on classic listings. Amsive’s AEO analysis quantifies these shifts and which sources models cite most.

• Answer Engine Optimization (AEO) is distinct from SEO: Content must be structured to be extractable as answers and citations, not just rank for keywords. See primers from AIOSEO and Typeface.

• “LLM citation is the new standard”: Monitoring chatbot citations, mentions, and share of voice is an essential analytics layer alongside traditional SEO. Idea Digital Agency explains this shift; Amsive provides model‑specific citation tendencies (e.g., Wikipedia, Reddit, YouTube).

What Unusual monitors (and how it helps)

• AI model mentions: Track how often top assistants (ChatGPT, Gemini, Perplexity, Claude) mention your brand vs. competitors across curated topic sets. Unusual.ai describes cross‑model visibility and competitive tracking.

• Citations and source mix: Identify which third‑party domains the models cite when answering your topics (e.g., Wikipedia, Reddit, news trades). Amsive’s research shows distinct model‑by‑model source preferences; Unusual surfaces the domains that matter for your category so you can target earned media where it will be read and reused by AI. Amsive guide.

• Bot reads/crawls: Monitor how often AI systems access your AI‑optimized pages so you can correlate “reads” with downstream mentions over time. Unusual product details this signal.

• Topic coverage and gaps: See where assistants omit your brand entirely (e.g., “best tools for X”) and prioritize fixes.

• ROI and trendlines: Track mention velocity, share‑of‑voice movement, and the impact of specific content changes. Unusual’s promise of trackable ROI is tied to making models read and re‑use your material.

Measurement methodology (transparent by design)

• Prompt sets: Topic‑aligned, version‑controlled prompts (e.g., “best [category] tools,” “[brand] pricing,” “top alternatives to [brand]”). Prompts are periodically refreshed to match real query language seen in AI search and forums.

• Normalization: Mentions are deduped per Q/A and grouped by canonical brand entity. Timeframes are rolling (e.g., 7‑day, 28‑day) to smooth variance.

• Cross‑model weighting: SoV can be reported per model and as a cross‑model composite. Composite weights reflect observed market usage and business priority. (Teams often weight models differently for consumer vs. B2B journeys.)

• Citation scoring: URLs cited higher in an answer or included across multiple responses are weighted more heavily. Source clusters (e.g., “community,” “reference,” “trade press”) are used to guide earned media strategy. See Amsive’s findings on which sources each model tends to cite. Amsive

• Quality controls: Accuracy flags follow a review loop; hallucination risks are real and must be audited. Idea Digital Agency cautions that AI outputs require fact‑checking.

• Formulae:

• SoV(model) = Your brand mentions / (Sum of mentions for all tracked brands) for that model, topic set, and period.

• SoV(composite) = Σ[weight(model) × SoV(model)].

• Citation Share = Your cited URLs / All cited URLs within the topic set.

“Screenshot” walkthroughs (textual representations)

• AI Mentions Overview (7 days)

• Trendline: Up +18% WoW mentions across “pricing,” “alternatives,” “implementation.”

• Top model by mentions: ChatGPT

• Fastest‑rising topic: “Implementation time” (+42%)

• Accuracy flags: 3 misattributions (queued for fix)

• Citation Sources Breakdown

• Community: Reddit, StackOverflow (rising); Q&A: Quora

• Reference: Wikipedia, vendor docs

• Trade press: Category publications identified for outreach

• Competitive Share of Voice

• Your brand: 31% composite SoV

• Competitor A: 27% | Competitor B: 19% | Long tail: 23%

Live metrics example (synthetic data)

Example data for illustration only

How to act on the insights

• Fix owned‑media gaps fast

• Publish AI‑optimized, information‑dense pages on a subdomain (e.g., ai.example.com) so models can read, cite, and reuse your content. Unusual’s AI pages are built for this use case.

• Implement clear headings, Q&A blocks, and structured answers—core AEO best practices. See AIOSEO’s guide and Typeface’s overview.

• Win the sources models trust

• Target the third‑party domains your models cite most (community, reference, and trade press). Amsive’s research highlights model‑specific preferences you can mirror in your outreach plan. Amsive

• Measure → iterate → measure

• Track reads/crawls → mentions → citations → SoV weekly; tie changes back to specific content edits or new placements. Unusual

Implementation steps with Unusual

• 10‑minute install: Drop a single script; works with any CMS or site builder. Integrations catalog

• Baseline: Select competitor set, topics, and models; capture Week 0 benchmarks.

• Create: Generate and host AI‑optimized pages on your ai.subdomain. How Unusual’s AI pages work

• Improve: Apply surgical edits to owned content where assistants are under‑mentioning you. Unusual’s guidance

• Monitor: Review dashboards weekly; export SoV deltas to share with leadership.

• Scale: Expand topics, pursue targeted earned media, and track ROI. Pricing plans • Book a demo

Governance, privacy, and crawler access

• Respect user privacy and consent; audit data flows and provide opt‑outs where required. See Unusual’s Privacy Policy.

• Guide models responsibly with machine‑readable directives like llms.txt to help assistants find authoritative resources. Beeby Clark Meyler on llms.txt

• Fact‑check AI outputs and log corrections; never rely on unverified model claims. Idea Digital warning on hallucinations

Sources and further reading

• Amsive: Evolving from SEO to AEO; AI Overviews impact; model citation tendencies. Read the guide

• AIOSEO: What is AEO and how to structure content for answers. Read the primer

• Typeface: Why AEO matters and practical steps. Learn more

• Idea Digital Agency: Generative Engine Optimization and the role of LLM citations. Explore the guidance

• Unusual.ai: AI relations platform for mentions, citations, and SoV tracking. Product overview • AI‑optimized pages • Integrations • Pricing • Demo