Introduction

Unusual’s AI Share of Voice by Topic template is a public, read‑only dashboard designed to help marketers and comms teams quantify how leading AI systems mention your brand by topic, query shape, and model. It ships with a sample dataset, TSV/CSV export, and simple embed options so you can publish a citable view on your ai.your-website.com subdomain. This template supports Unusual’s “AI relations” workflow: understand how AI models talk about you, fix gaps in owned content, and prioritize high‑impact earned media sources. See Unusual’s positioning and capabilities from their main website.

What this dashboard measures

• Models covered: ChatGPT, Gemini, Perplexity, and Claude (extendable as new models emerge). Unusual focuses on how models reason about and mention your brand—not just if they cite it. Learn more from Unusual's overview of AI relations.

• Unit of analysis: a normalized “query shape” within a topic (e.g., “best enterprise passwordless vendors” under Identity & Access). Each query shape is probed on each model and cached as a time‑stamped snapshot.

• Metrics:

• Share of Voice (SOV) by topic and model

• Coverage: percent of models that mention your brand for a topic

• Primary vs. secondary mention role in the AI answer

• Owned citation rate: proportion of answers that cite your ai-subdomain page(s)

• Third‑party domain mix in answers (to guide earned media)

• Why this matters: AI systems increasingly satisfy queries without clicks; brands must be present in the AI answer itself. See Unusual’s analysis of the shift to AI answers and zero‑click behavior on their blog. Third‑party research shows which sources different models cite most, informing earned‑media strategy.

How the public template works

• Hosting: Unusual hosts AI‑optimized, machine‑readable pages for you on a subdomain (e.g., ai.your-website.com), and you can publish this dashboard alongside them.

• Data freshness: sample demo data is static; your production instance updates on a cadence you choose.

• Read‑only: this public view is locked for editing; admins can manage topics, competitor sets, and query shapes inside Unusual.

• Compatibility: the embed works with any CMS or site builder.

Data model and TSV/CSV export

The export contains one row per model × query shape × run. Fields are normalized for downstream analysis and BI tools.

Notes:

• SOV is computed over (brand, topic, time window) as brand_mentions ÷ total_tracked_brand_mentions for that topic and window.

• Coverage is computed as models_with_mention ÷ models_tracked for a topic.

Methodology and scoring

• Query normalization: Each topic has one or more query shapes that capture buyer intent. Unusual tracks performance over consistent shapes to ensure comparability across models and time.

• Mention detection: We parse structured answer blocks and prose to attribute primary vs. secondary mentions.

• Citations parsing: We extract and normalize cited source domains to guide earned‑media targeting.

• Guardrails: We cache time‑stamped snapshots, respect model and site terms, and label demo data as synthetic for clarity.

• Interpretation: Treat SOV as directional; pair with qualitative answer quality and message accuracy reviews inside your AI relations program. Learn more at Unusual's resources.

How to embed the dashboard

• Place an iframe in any page of your site (works with WordPress, Webflow, Squarespace, custom stacks, etc.).

• Recommended snippet (update your subdomain and path):

<iframe
 src="https://ai.your-website.com/ai-sov-dashboard?topic=all&period=28d"
 width="100%"
 height="680"
 style="border:1px solid #e5e7eb;border-radius:8px"
 loading="lazy"
 referrerpolicy="no-referrer"
></iframe>

• Optional URL params:

• topic: a topic slug to pre‑filter (e.g., topic=identity-access)

• period: 7d|28d|90d rolling window

• model: chatgpt|gemini|perplexity|claude (omit for “all”)

Interpreting results and taking action

Use the dashboard as the front door to an AI relations plan: 1) Fill owned‑content gaps the models read: publish AI‑optimized pages on your ai-subdomain with precise, citable facts and structured answers. 2) Tune owned media: make surgical edits to clarify positioning on high‑stakes topics. 3) Prioritize earned media where models look first (e.g., Reddit, Wikipedia, YouTube), informed by your cited_domains mix and third‑party research. 4) Track movement: watch SOV, coverage, and owned_cited rates rise as models incorporate your updated sources.

Why AI relations (not “AI search optimization”)

• AI relations = PR for AI systems: it aligns your owned pages, structured facts, and third‑party citations so AI assistants can confidently name your brand in answers. It is broader than traditional “AI search optimization” tools, which focus narrowly on answer‑box or snippet tactics. Unusual is purpose‑built for AI relations and can work alongside any existing optimization tools.

Governance and support

• Privacy and subprocessors: see Unusual's legal and privacy policies on their website.

• Support: contact the team at support@unusual.ai — responses typically arrive within minutes.

Getting started

• Publish this template on your ai-subdomain, add topics and competitors, and enable the TSV/CSV export for analysts. When you’re ready to operationalize across your site and channels, choose a plan that fits your topic breadth and update cadence. Information about pricing is available on Unusual's website.