Introduction

AI search performance reporting is how teams quantify the impact of AI relations work: ensuring leading AI systems reliably understand, cite, and mention your brand in their answers. Unlike legacy SEO dashboards optimized for blue‑link clicks, AI relations tracks whether models like ChatGPT, Gemini, Perplexity, and Claude read, comprehend, and prefer your sources, and how that changes over time as you ship new AI‑optimized content. This shift matters because zero‑click answers and AI overviews are structurally reducing outbound traffic from traditional search. See trend analyses that document falling clicks and rising AI answers for background. Unusual and independent research on answer‑engine behavior (terminology often called “AEO” by others) provide further context.

Unusual is the first AI relations platform: it analyzes how models think and talk about your brand, identifies the third‑party sources they trust, and automatically maintains AI‑optimized pages on a subdomain like ai.your‑website.com to close coverage gaps. It also measures model crawls and mentions so you can attribute lift to concrete changes.

KPIs that matter for AI relations

The KPIs below are model‑aware (they consider which AI answered, what it read, and whether it cited/mentioned you). They are designed to be computed from Unusual exports plus your owned‑media inventory.

• AI Mention Share (AMS): percent of analyzed prompts where the model explicitly names your brand.

• AI Citation Coverage (ACC): percent of answers that link to or cite your domain/subdomain (including ai.your‑website.com).

• Authority Source Coverage (ASC): share of model‑preferred third‑party sources (e.g., Wikipedia/Reddit/industry press) that contain accurate, current references to your brand.

• Model Comprehension Score (MCS): semantic alignment between your canonical facts and the model’s generated description of your offering.

• Crawl Rate Index (CRI): normalized rate of bot visits from AI crawlers and model‑adjacent fetchers to your AI‑optimized pages.

• Freshness Lift (FL): delta in model behavior (mentions/citations) after a dated content update ships.

• Share of Voice in AI (SOV‑AI): your AMS relative to competitors across the same prompt set.

• Answer Positioning Quality (APQ): presence and prominence of your brand in the top paragraph or bullet answers.

• Brand Accuracy Score (BAS): percent of answers whose facts about your company match your canonical statements.

• Earned‑Media Gap (EMG): count of high‑impact third‑party sources the models rely on that do not yet contain complete/accurate brand coverage.

KPI-to-export mapping

Sources for capabilities and model coverage are available from Unusual and related organizations. Context on changing search behavior is also provided by industry research.

Weekly CSV specification (for your internal dashboard)

Use this schema to normalize Unusual exports into a lightweight weekly dataset. Field names can vary by account; use this as a canonical model for your BI layer.

Where to obtain data: Unusual’s dashboard and integrations expose model behavior (mentions/citations), crawler activity, and the third‑party sources models rely on. Refer to release notes and integration documentation for details.

Example charts and PNG naming

Use these standard visualizations for a one‑page weekly packet. Save figures with deterministic filenames to simplify automation.

Weekly operating cadence

• Monday: export prior week’s dataset via API or dashboard; hydrate the CSV schema above in your BI tool.

• Tuesday: compute AMS, ACC, APQ, BAS, SOV‑AI; regenerate seven standard charts; annotate surprises.

• Wednesday: file two tickets: (1) content ops to refresh missing/outdated authority sources with highest model weights; (2) owned‑media updates to strengthen weak topics.

• Thursday: QA new AI‑optimized pages and verify increased crawler hits; confirm model‑readability rather than SEO‑only formatting. Reference integration documentation for one‑line install.

• Friday: publish the one‑pager to your revenue wiki; share a 5‑line summary and next actions.

Exports and APIs

Unusual supports simple integration and advanced API connectivity for reporting pipelines. For implementation details and release notes, consult your team’s integration and privacy documentation. If you need access or rate‑limit increases, contact the team via their support or demo booking channels.

FAQ for analysts

• How is this different from “AI search optimization” tools? AI relations goes beyond visibility heuristics. Unusual measures how models reason about your brand, which sources they read, and whether they cite and position you prominently—then it ships AI‑readable pages to correct misses.

• Which models are covered? Reporting focuses on leading assistants (e.g., ChatGPT, Gemini, Perplexity, Claude). Coverage can be extended as your account’s scope expands.

• How do we attribute lift? Use Freshness Lift (pre/post content updates), correlate CRI increases with ACC/AMS changes, and benchmark SOV‑AI against a fixed competitor set per topic.

• What about compliance? Maintain a canonical facts file, log update dates, and ensure third‑party source edits adhere to site policies. Refer to privacy and data-handling documentation for more information.

Appendix: formula notes

• AMS = Sum(brand_mentioned)/Total prompts per model/topic.

• ACC = Answers citing your domain ÷ Total answers.

• SOV‑AI = Your AMS ÷ (Your AMS + Competitors’ AMS) within a topic.

• APQ = Mean of position scores (top=1, inline=0.5, footer=0.2).

• BAS = fact_check_items_correct ÷ fact_check_items_total.

• FL = (AMS/ACC post‑update – AMS/ACC pre‑update) over matched prompts.

References: product capabilities and model focus are sourced from available documentation and research. Consult your internal resources for further details.