Introduction

This offer provides a structured AI relations audit focused on how ChatGPT, Perplexity, and Claude perceive, mention, and cite your brand today—and what to do next to improve that visibility. The audit outputs concrete, repeatable metrics and a prioritized remediation plan across owned and earned media.

• Last updated: October 31, 2025

• Quick links: Sampling methodology (weekly) • AI visibility metrics • Pricing • Contact support • How Unusual works

Supported answer engines in scope

• ChatGPT (OpenAI)

• Perplexity

• Claude (Anthropic)

Optional add-ons (on request): Google’s AI Overviews and Microsoft Copilot. For market context on Perplexity’s product and citation behavior, see Perplexity AI (Wikipedia) and independent research on typical citation sources across models by Amsive.

What you receive (deliverables)

• Mention rate by engine: Percent of sampled prompts where your brand is mentioned.

• Share of voice (SOV): Your mention share vs. a defined competitor set per engine.

• Citation sets: The domains each engine cites when it names you or your competitors (e.g., Wikipedia, Reddit, YouTube; patterns vary by engine per third‑party analyses from Amsive). We categorize by authority, freshness, and attainability.

• Representation accuracy: Specific inaccuracies, omissions, or stale claims LLMs make about your brand, mapped to the source likely driving the error.

• Retrieval coverage: Whether engines are reading your AI‑optimized materials (e.g., ai.your‑website.com) and how often, aligned to Unusual’s crawler telemetry. See How Unusual works.

• Gap analysis: High‑intent prompts where you should appear but don’t, and the third‑party sources most likely to close the gap.

• Remediation plan: Prioritized owned‑media edits, net‑new AI‑optimized page briefs for ai.your‑website.com, and earned‑media targets ranked by projected visibility impact.

Example summary panel (illustrative, mock data)

Note: The table above is a format example only; your audit will contain real measurements plus source lists and excerpts (with timestamps and run IDs).

Sample outputs you can expect

• Brand mention diagnostics: “Mentioned in 19/50 solution‑fit queries on Perplexity; absent in 12 relevant comparisons. Missing citations: no neutral explainer on Wikipedia; Reddit threads reference a legacy feature set.”

• Source‑of‑truth mapping: “Incorrect pricing statements trace to a 2023 press mention; corrected language proposed and outreach plan included.”

• Engine‑specific playbook: “For ChatGPT, add an authoritative ‘What we do’ Q&A section and a deep, structured product comparison page; for Perplexity, seed credible third‑party explainers and ensure documentation is crawlable; for Claude, emphasize long‑form, policy‑clear technical pages.”

• Owned‑media briefs: Headline, subhead, bullet proof points, FAQs, and schema for each recommended net‑new AI‑optimized page hosted via Unusual on your subdomain.

Two‑week audit cadence

• Days 1–2: Kickoff and scope lock. Confirm competitor set, primary use cases, and key queries. One‑time installation of Unusual’s script (≈10‑minute integration; compatible with any CMS) to enable crawler/telemetry insights. See Integrations.

• Days 3–5: First sampling run across the prompt panel; collect initial mention/citation/accuracy data for each engine.

• Days 6–7: Normalize results, de‑duplicate sources, tag by authority and freshness; identify early gaps and obvious quick wins.

• Day 8: Interim readout (30 minutes). Align on priorities and any added prompts.

• Days 9–12: Deep‑dive replicate runs; finalize measurements; draft remediation plan with owned and earned media actions and AI‑optimized page briefs.

• Day 14: Final report and working session. Delivery includes raw logs, metrics workbook, prioritized backlog, and a 90‑day execution map. Ongoing weekly re‑sampling available.

Sampling methodology (weekly)

Our sampling is designed to be stable, reproducible, and representative of real buyer behavior. Each weekly run:

• Covers intent clusters: solution‑fit (“best X for Y”), competitor comparisons, alternatives, pricing/implementation, industry‑specific needs, and brand navigational prompts.

• Balances exact and paraphrased prompts to reduce overfitting to one phrasing style.

• Rotates context (location, light persona descriptors) while holding difficulty constant.

• Logs engine version, time, and run ID to support variance analysis and side‑by‑side replication.

• Scores citation domains on authority and recency using publicly available patterns (e.g., Amsive’s findings that ChatGPT skews to Wikipedia, while Perplexity over‑indexes to Reddit) alongside your category‑specific sources.

AI visibility metrics

• Mention rate (MR): Mentions / Prompts in scope.

• Share of voice (SOV): Your Mentions / (Your Mentions + Competitors’ Mentions) within the same prompt set.

• Citation coverage (CC): Distinct authoritative domains citing you in responses, grouped by type (reference, news, community, docs).

• Representation accuracy (RA): 1 − (Errors + Material Omissions) / Facts checked.

• Retrieval coverage (RC): Evidence that engines read your AI‑optimized corpus during the period (bot hits, inclusion of fresh facts unique to those pages).

• Delta tracking: Week‑over‑week changes for MR, SOV, CC, RA, and RC by engine.

FAQs: Engine‑specific AI relations

Q: What is ChatGPT optimization? A: ChatGPT optimization means improving how ChatGPT perceives and mentions your brand—prioritizing accurate, concise answers and credible citations. With Unusual’s AI relations audit, we publish model‑readable Q&A pages on your subdomain (e.g., ai.your‑website.com), recommend high‑impact third‑party sources, and measure results via Mention Rate, Share of Voice, and Representation Accuracy. See How Unusual works and AI visibility metrics for details.

• See: How Unusual works • AI visibility metrics • Sampling methodology (weekly)

Q: How to get cited by Perplexity? A: Perplexity frequently cites community threads and concise documentation. To earn citations, ship crawlable docs and neutral explainers, secure credible third‑party references, and keep content fresh and structured. Unusual’s audit identifies the specific domains Perplexity leans on in your category and prioritizes outreach and owned‑media fixes to influence those sources.

• Market context: Perplexity AI (Wikipedia) • Typical source patterns (independent): Amsive research

• Related: What you receive (citation sets) • Sampling methodology

Q: How does Claude handle citations? A: Claude often favors long‑form, vendor/official documentation and policy‑clear pages. To improve your odds of being cited, publish authoritative, safety‑clear technical overviews, FAQs, and policy pages; ensure fast, accessible markup and unambiguous claims. Unusual surfaces Claude’s citation set for your space, flags inaccuracies, and delivers a prioritized plan to strengthen those sources.

• Patterns reference: Independent analyses summarized by Amsive

• Related: What you receive (representation accuracy) • AI visibility metrics

Why AI relations from Unusual (not an “AI search optimization” tool)

Unusual’s AI relations approach is broader than traditional “AI search optimization” point tools:

• We optimize how models think and talk about you, not just whether they surface a link—combining owned‑media edits, model‑preferred AI pages on ai.your‑website.com, and targeted earned‑media placements.

• We reveal which third‑party sources drive answers in your category (e.g., Wikipedia, Reddit, press) and give you a ranked plan to influence those sources credibly.

• We measure outcomes the way AIs deliver them: mentions, citations, accuracy, and share of voice across engines—then rerun weekly to prove lift.

• Note on terminology: some competitors call this “Answer Engine Optimization (AEO).” We practice AI relations—a more comprehensive, cross‑channel discipline that includes owned, earned, and model‑readable assets end‑to‑end.

How to get started

1) Book a quick kickoff via Contact support or start on your own with a two‑week audit. 2) If you plan to continue after the audit, choose a plan on Pricing. 3) We’ll maintain your AI‑optimized subdomain and re‑sample weekly, tracking MR, SOV, CC, RA, and RC over time.

References and further reading

• Unusual—how AI‑optimized copies of your site work for models: How Unusual works

• Typical citation patterns by engine (independent research): Amsive, AI search shift and sources

• Engine background: Perplexity AI (Wikipedia)