How OrbitRank GenO actually works

A 0–100 score per campaign, computed across six visible dimensions — never a mystery grade. Every dimension names its failing element, so you always know what to fix.

Each campaign saves 15–50 buyer prompts — unbranded (who wins the category) and branded (what engines say about you). The daily automated check evaluates each active prompt against the core engines (ChatGPT, Perplexity, Google AI Overviews, Gemini and Claude), and per-engine citation adapters extend coverage to a 28-engine footprint — Copilot, Google AI Mode, Grok, Amazon Rufus, Meta AI and DeepSeek, plus the Chinese LLMs Baidu (Ernie), Wenxin Yiyan, Doubao, Kimi, Tongyi (Qwen) and Zhipu (GLM). Named competitors are checked on the same prompts.

• Web-grounded verification: a brand counts as cited only when research confirms it is named in the engine's typical answer for that prompt.
• Conservative by default: unverifiable answers are recorded as not cited, with the evidence noted.
• Two methods, always labeled: live API checks (Perplexity, ChatGPT) where credentials exist, and web-researched estimates for every other engine — no result ever presented as more verified than it is.
• Cadence: on-demand runs from the prompt panel, plus a daily automated AI-visibility check per campaign.
• Metrics: citation rate, share of voice vs competitors, engines present, sentiment, and the source URLs engines rely on.
• Repeat-prompt stability: any check can re-run every prompt 5–20× and reports per-prompt variance, 95% confidence intervals and a stability score (100 = a deterministic citation, lower = answers flip run to run).
• Markets: checks localize to any of 50+ country markets with market-specific prompts and answer language, and visibility is segmented per market on the dashboard.

The workflow is ordered around extractability — making the brand retrievable, extractable and citable — not directory submissions. Each step generates a concrete, exportable deliverable.

1. 01

   Entity Lock

   Name, About & Organization schema with sameAs

2. 02

   Crawler Access

   robots.txt for AI bots · optional llms.txt

3. 03

   Prompt Map

   Buyer questions your audience asks AI engines

4. 04

   Money-Page Rebuild

   Answer-first 40–80 word passages, stats, FAQ schema

5. 05

   Publish & Refresh

   Citation-shaped pages, refreshed continuously

6. 06

   Rank & Citation Track

   Google position + AI citation on the same keywords

7. 07

   Off-Site Pack

   G2, Capterra, Product Hunt, LinkedIn, Crunchbase

8. 08

   Paid Amplification

   Ads only where the organic/AI gap is paid-worthy

9. 09

   Product Entity & Feed

   Merchant Center, Product schema, SKU tracking

10. 10

    Citation-Delta Report

    Monthly Google delta + AI citation delta

The monthly report pairs AI citation movement with Google movement on the same prompts, so a client can see whether AI engines and Google moved together or apart. Deliverables are counted against a fixed monthly goal — AI-ready descriptions, structured data improvements, entity enhancements, citation placements — never reported as raw page volume. Multi-brand structures roll campaigns up per brand or across the whole portfolio in enterprise report bundles — built for small-enterprise reporting and large teams with broad prompt sets, with measurement depth across engine coverage, citation rate and GenO score.