How We Assess AI Visibility
When a Singapore business asks us "how visible am I to AI models?" we do not guess. We run a structured audit across five dimensions that collectively determine whether ChatGPT, Claude, Perplexity, and other AI tools can find, understand, and recommend your business.
This framework was developed from auditing dozens of Singapore businesses and correlating their scores with actual AI citation frequency. We publish the framework because we believe business owners deserve to understand where they stand.
Dimension 1: Structured Data
Structured data is the foundation of AI visibility. It includes schema markup (JSON-LD) on your website, consistent business listings across directories, and machine-readable data about your services, location, hours, and offerings.
What we check: does your website have LocalBusiness schema? Product or Service schema? FAQ schema? Are your Google Business Profile, Facebook, and directory listings consistent? Is there structured pricing, availability, or feature data that AI models can parse?
Why it matters: AI models can parse structured data with near-perfect accuracy. Unstructured text requires inference, which introduces errors. A business with comprehensive schema markup gives AI models concrete facts to cite rather than guesses.
Scoring: we rate structured data completeness, accuracy, and consistency on a five-point scale. Most Singapore SMBs score 1 to 2 out of 5 because they have minimal schema markup and inconsistent directory listings.
Dimension 2: Content Authority
Content authority measures whether your website positions your business as an expert in your field. AI models preferentially cite sources that demonstrate depth of knowledge on a topic.
What we check: do you have comprehensive content about your services, industry, and customer questions? Is the content well-structured with clear headings and logical organisation? Does it cover topics thoroughly enough for AI models to extract specific answers? Is it regularly updated?
Why it matters: a dental clinic with a single "Services" page listing treatment names has low content authority. A dental clinic with detailed pages on each treatment, a blog covering common dental questions, and educational content about dental health has high content authority. AI models cite the second clinic because they can extract more specific and useful information from it.
Scoring: we evaluate content depth, breadth, structure, and freshness. Content authority is where most businesses have the most room for improvement.
Dimension 3: Entity Recognition
Entity recognition measures whether AI models recognise your business as a distinct entity rather than a generic mention. Strong entity recognition means the AI knows your business name, what you do, where you operate, and what distinguishes you from competitors.
What we check: does the AI recognise your business by name when asked directly? Is there a Google Knowledge Panel for your business? Is your business listed on Wikidata? Is your brand name consistent across all online mentions? Are there enough independent mentions for AI models to establish you as a real entity?
Why it matters: if you ask ChatGPT about a business and it responds with "I don't have specific information about [business name]," that business has low entity recognition. AI cannot recommend what it does not recognise.
Scoring: we test entity recognition across multiple AI models and score based on recognition rate, accuracy, and detail of responses.
Dimension 4: Citation Signals
Citation signals are the external validations that make AI models confident in recommending your business. These include reviews, press coverage, industry mentions, and third-party endorsements.
What we check: Google review count and average rating, reviews on industry-specific platforms, mentions on news sites and industry publications, backlinks from authoritative domains, and social proof across platforms.
Why it matters: AI models use third-party validation as a confidence check. A business recommended by multiple independent sources is more likely to be cited than one with no external validation. This is similar to how Google uses backlinks, but AI models also weigh review sentiment and quantity.
Scoring: we assess review volume, review quality, media mentions, and authoritative backlinks.
Dimension 5: Technical Accessibility
Technical accessibility measures whether AI crawlers can actually access and read your website content. A beautifully designed website that blocks AI crawlers or hides content behind JavaScript is invisible to AI models.
What we check: can AI crawlers access your content (robots.txt configuration)? Do you have an llms.txt file? Is your site architecture clean and crawlable? Is important content rendered in HTML (not hidden in JavaScript or images)? Does your site load quickly?
Why it matters: none of the other dimensions matter if AI models cannot read your content. Technical accessibility is the prerequisite for everything else.
Our llms.txt guide and MCP server explainer cover the most advanced technical accessibility measures.
Using the Framework
You can run a basic self-assessment using this framework. For each dimension, rate your business on a 1 to 5 scale. This gives you a 25-point AI visibility score and, more importantly, shows you which dimensions need the most work.
Most Singapore SMBs score between 5 and 10 out of 25. The businesses that score above 15 are consistently cited in AI recommendations. The gap between where most businesses are and where they need to be represents the opportunity for early movers.
For a detailed guide on how to improve each dimension, see our AEO guide. For a case study showing how one brand moved from a low score to consistent AI citations, see our AEO case study.
Talk to Us
Chat with us on WhatsApp to request a free AI visibility audit. We reply within one Singapore business day.
Ready to get started?
Chat with the Swop Labs team on WhatsApp. We reply within one Singapore business day.
Chat on WhatsApp