The AI Undercurrent

Whether AI systems can physically reach, crawl and read your website. Access is a gate: if AI can’t get in, nothing else you do matters.

• L0.1AI Crawler AccessCriticalWhether robots.txt blocks any of the known AI crawlers, such as GPTBot, ClaudeBot and PerplexityBot.
• L0.2Google Index StatusWhether your key pages are indexed and discoverable.
• L0.3SSR / Pre-renderingWhether your content is in the raw HTML, not only after JavaScript runs.
• L0.4llms.txtWhether an llms.txt file gives AI crawlers a structured overview of your business.

What good looks likeAI crawlers can reach every important page, the content is in the raw HTML, and an llms.txt file gives AI a structured overview of your business.

Whether AI models correctly understand your brand, positioning, services, locations and market.

• L1.1Core Positioning AccuracyCriticalWhen AI describes your brand, does it get the positioning right?
• L1.3ICP CoverageDoes AI know who your customers are?
• L1.6Geographic AccuracyDoes AI know everywhere you operate?
• L1.10Notability SignalsIs there enough evidence (reviews, press, awards) that your brand is notable?

What good looks likeAsk AI “Tell me about [your brand]” and it accurately describes your positioning, services, locations and audience, and treats you as an established entity.

Whether AI echoes your brand messaging or substitutes its own. Most brands discover AI is writing their story in words they’d never choose.

• L2.1Echo Phrase AlignmentDo your key brand phrases appear in AI responses?
• L2.4Sentiment Bias ScoreWhat’s the emotional tone when AI mentions you?
• L2.6Canonical Language AdoptionHave you defined the exact sentences you want AI to use, and are they getting through?

What good looks likeAI repeats your own positioning phrases, with a positive tone, because your content gives it clear language to echo.

Whether independent, trusted sources back up what you say. Below the corroboration threshold AI hedges (“claims to be”); above it, AI asserts (“is”). The flip happens at two to three independent sources per claim.

• L3.1Superfeeder PresenceCriticalAre you present on the platforms AI draws from most, such as YouTube, Reddit, review sites and industry publications?
• L3.3Citation Portfolio BreadthHow many of the 7 AI platforms cite you?
• L3.8Non-Branded Share of VoiceDo you appear when people ask about your category without naming you?

What good looks likeAll 7 AI platforms cite you, and at least two or three independent sources confirm each key claim, so AI states facts about you instead of hedging.

Whether your content is formatted so AI can extract and quote it. AI pulls specific passages and facts; pages that lead with marketing copy give it nothing to use. This is where most brands fail, and the fastest layer to fix.

• L4.1Fan-Out CoverageCriticalWhen AI breaks a query into sub-questions, can your site answer them?
• L4.3Answer-First StructureDoes your content lead with the answer?
• L4.4FAQ ExtractabilityIs there FAQ content, with schema, that AI can lift as question-and-answer pairs?
• L4.5Schema ImplementationDoes structured data describe your entity, locations, ratings and services?

What good looks likePages open with facts, cover the sub-questions buyers ask, and carry FAQ and organization schema that AI can extract directly.

Whether what AI says about you is accurate, consistent and fair. If your own site contradicts itself, AI picks the wrong fact or stops citing you.

• L5.1Hallucination FrequencyCriticalHow often does AI state incorrect facts about you?
• L5.2Contradiction RiskDoes your site say different things on different pages?
• L5.7Transparency SignalsAre pricing and policies clear and visible?
• L5.9Negative Narrative CounterWhen AI surfaces negative themes, does your site address them directly?

What good looks likeAI states accurate facts consistently, your site has no internal contradictions, and common concerns have a clear answer on your site.