Ask ChatGPT to recommend a plumber, a dentist, or a web designer in your city. Odds are it will name a few businesses — and leave out most of the legitimate, well-reviewed companies that have operated in that market for years. The businesses that get cited aren't necessarily the best. They're the ones whose digital footprint is structured in a way AI engines can read, trust, and cite. Understanding why that happens — and how to fix it — is what AI visibility is all about.
Training Data vs. Live Citations: Two Different Problems
The first thing to understand is that not all AI tools work the same way. ChatGPT's core model was trained on a massive snapshot of the web up to a specific knowledge cutoff date. After that date, new information about your business — a new address, a new service, a new website — simply does not exist in the model's memory unless it has been updated through a newer training run. This is why some businesses that were well-documented years ago still show up in ChatGPT answers, while newer businesses with great websites are invisible.
Perplexity AI and Google's AI Overviews work differently. They pull live web content at query time, then synthesize an answer. These tools do read your website — but only if they're allowed to, and only if they can parse what they find. A beautiful website that's blocked to AI crawlers, or a site with no structured data, gives these tools nothing to work with even if they can technically access it.
This distinction matters because the fix for each problem is different. For ChatGPT's knowledge cutoff problem, the answer is entity establishment — building a consistent, authoritative record of your business across Wikipedia-adjacent sources, business directories, and structured data. For live-citation tools, the answer is structured markup and open crawler access.
The Role of Structured Data
Structured data — specifically Schema.org markup embedded in your website's HTML — is the language AI engines use to understand what your business is, where it operates, what it offers, and who it serves. Without it, an AI crawler reads your site the way a stranger reads a handwritten sign in a foreign language: they can see something is there, but they can't tell what it means with confidence.
When your site includes LocalBusiness schema with your exact name, address, phone number, hours of operation, and service area, an AI engine can read that as structured, verifiable data — not just text on a page. When you add FAQPage schema with the actual questions your customers ask, you're handing AI engines pre-formatted answers they can cite almost verbatim.
Most small business websites have none of this. Their contact page has an address written in a text block. Their services page is a paragraph of marketing copy. Their FAQ lives as an accordion widget with no machine-readable markup behind it. To a human visitor, it looks fine. To an AI engine, it's noise.
Key insight: "AI engines don't reward good design or compelling copy — they reward structural clarity. The business with a plain-looking site and complete Schema markup will consistently outperform the beautiful site with no structured data in AI-generated answers."
Why Blocking AI Crawlers Is Fatal
In the early days of AI, some businesses and publishers blocked AI crawlers in their robots.txt file — the configuration file that tells web robots which pages they're allowed to access. The thinking was: "I don't want AI tools scraping my content for free." The unintended consequence is that those businesses are now invisible to AI citation engines like Perplexity and Google AI Overviews.
Common AI crawler user agents include GPTBot (OpenAI), PerplexityBot, Claude-Web (Anthropic), and Google-Extended. If your robots.txt has a blanket Disallow: / for any of these bots, those AI engines will not read your site — period. They will not cite you, reference you, or recommend you, no matter how good your content or reviews are.
The fix is straightforward: audit your robots.txt file and ensure AI crawlers have access to your key pages. This is one of the first things I check in every AI Visibility Audit.
NAP Consistency: The Signal AI Uses to Trust You
NAP stands for Name, Address, Phone — the three core data points that define a business's identity online. AI engines that pull from live directories, Google Business Profile, Yelp, Bing Places, Apple Maps, and dozens of other sources cross-reference these data points to determine whether a business is real, active, and trustworthy.
If your Google Business Profile says "US WebSites" but Yelp says "US Websites LLC" and your website footer says "US Web Sites," those look like three different entities. AI models learn from patterns, and inconsistency reads as uncertainty. Businesses with perfectly consistent NAP data across 50+ directories carry far more entity authority than businesses with mismatched or missing listings.
Common NAP consistency problems I find during audits:
- Old phone numbers or addresses still live on directories from years ago
- Slightly different business name formatting across platforms
- Duplicate listings on the same directory with conflicting information
- Missing listings on high-authority directories (Apple Maps, BBB, Foursquare)
- Website footer that doesn't exactly match the Google Business Profile address format
The AI Platform Ecosystem: Every Engine Moves at a Different Speed
One of the most common misconceptions in GEO is treating all AI platforms as interchangeable. They're not. Each has a fundamentally different ingestion model — meaning the time between you making changes and those changes appearing in AI-generated answers varies significantly by platform.
| Platform | Initial Results | Full Optimisation | Key Factors |
|---|---|---|---|
| ChatGPT | 2–4 weeks | 3–5 months | Periodic training updates, authority |
| Google AI Overview | 1–3 weeks | 2–4 months | Traditional SEO signals, E-A-T |
| Perplexity | 1–2 weeks | 2–3 months | Real-time indexing, source diversity |
| Claude | 3–5 weeks | 4–6 months | Content quality, factual accuracy |
Perplexity is the fastest to reflect changes — its real-time indexing means a well-structured update can show up in answers within days. ChatGPT's core model is the slowest because it depends on periodic training runs, not live crawling. Google AI Overviews sits in the middle, inheriting much of traditional Google's indexing speed but weighted by E-A-T signals. Claude prioritizes content quality and factual accuracy above all else, which means thin or repetitive content is unlikely to gain traction regardless of when it's indexed.
A proper GEO strategy accounts for all four. Quick wins on Perplexity and Google AI Overviews buy you early visibility while the longer-arc work on ChatGPT and Claude builds your permanent entity authority. Without a plan that acknowledges these differences, you're optimizing for an average that doesn't actually exist.
Share of Voice: The Metric That Actually Matters
Traditional SEO metrics — keyword rankings, domain authority, backlink counts — tell you how you're doing in Google's index. They tell you almost nothing about how you're doing in AI-generated answers. The metric that matters in AI visibility is share of voice: out of all the times someone asks an AI engine a question relevant to your business category and geography, what percentage of responses mention your business?
A healthy local business in a competitive market should aim to appear in 30–50% of relevant AI queries. Most small businesses I audit are at 0–5%. The gap between those numbers is exactly what Generative Engine Optimization (GEO) closes — through structured data, entity establishment, NAP cleanup, crawler access, and content that directly answers the questions AI engines are being asked.
The first step to improving your share of voice is knowing where you stand. That's exactly what a free AI Visibility Audit shows you: your current citation rate, which AI engines cite you, where your structured data is missing, and what your competitors are doing that you aren't.
Find Out If ChatGPT Knows You Exist
Free AI Visibility Audit — see your citation rate, structured data gaps, and NAP consistency score.
Get My Free Audit → View GEO Services