The Fundamental Shift: From Ranking to Answering

The key difference

Traditional search returns a ranked list and lets the user decide. AI search generates an answer and makes the decision for them. The question is no longer "which pages are most relevant?" It is "which sources can I trust enough to cite directly?"

Google's traditional algorithm has been refined over two decades to rank pages by relevance and authority. It is fundamentally a sorting problem. AI search is a different problem. When ChatGPT tells someone "I recommend calling XYZ Plumbing in Manchester", it has made a specific claim about a real business. If that claim is wrong, the platform looks unreliable. So AI systems are far more conservative about what they cite.

That conservatism is what creates the gap between Google visibility and AI visibility. And it is why closing that gap requires specific, deliberate action.

How Each Major Platform Works

ChatGPT
OAI-SearchBot
Combines a training knowledge base with real-time Bing search results. For business queries, it searches Bing, synthesises results, and cites sources. Bing's index and your structured data are the primary levers. Allowing OAI-SearchBot in robots.txt is essential.
Perplexity
PerplexityBot
Built as a search engine with AI synthesis on top. Crawls independently and shows citations more prominently. Rewards clear, direct content that answers specific questions. Favours content with clear authorship and publication dates.
Google AI Overviews
Googlebot
Draws on the full Google index plus the Knowledge Graph, which aggregates structured entity data from across the web. Schema markup, Google Business Profile, and verified entity signals all feed directly into the Knowledge Graph.

The Process: How an AI Decides What to Cite

1

Query intent classification

The AI identifies what kind of answer is needed. "Best accountant in Leeds" requires a business recommendation with location. The intent determines which sources are eligible. For local business queries, entities with verified location data rank much higher as candidates.

2

Source candidate retrieval

The AI retrieves candidates from its index. For ChatGPT and Copilot, this is Bing's index. For Google AI Overviews, it is Google's own index. Businesses that are not in these indexes, or that have blocked crawlers, do not enter the candidate pool at all.

3

Entity verification

This is where most small businesses fall out of the process. The AI cross-checks candidates against structured data signals: is there schema markup declaring a verified business identity? Does it match data in Companies House, LinkedIn, or Google Business Profile? Businesses that fail entity verification are not cited.

4

Relevance and authority scoring

Verified candidates are scored for how well they match the query and how authoritative they appear. Authority signals include off-site mentions, inbound links, content depth, and E-E-A-T signals. A verified entity with strong off-site presence beats a verified entity with none.

5

Answer generation and citation

The AI generates its answer using the highest-scoring verified sources. The businesses cited are the most authoritative among those that passed entity verification. That is why fixing your structured data matters so much. It is the gate you have to get through before authority even becomes relevant.

The Signals That Matter Most

Based on published research from Semrush, BrightEdge, and WPRiders, here is the relative impact of key signals for local business citation in AI search.

Schema markup (structured data)
High
Google Business Profile completeness
High
FAQ content with direct answers
High
Off-site mentions (Reddit, directories)
Med
External reviews
Med
Traditional backlinks
Low
Keyword density on page
Low

Most small businesses have spent years optimising the low-impact signals and have done almost nothing about the high-impact ones.

How Do You Score Against These Signals?

The free AI Visibility Snapshot checks your homepage across seven dimensions covering structured data, content, and accessibility. You will know exactly where you stand within 48 working hours.

Get Free Snapshot →

The Three-Layer Model for AI Visibility

Layer 1: Technical accessibility. Can AI systems actually read your site? Robots.txt, sitemap, page speed, crawl accessibility. This is the prerequisite. Nothing else matters if crawlers cannot get in.

Layer 2: Verified identity. Do AI systems know who you are? Schema markup, Google Business Profile, external entity verification, consistent NAP data. This is the gate. Pass it and you are in the candidate pool.

Layer 3: Authority and relevance. Among verified candidates, why should AI systems cite you over someone else? Content quality, off-site mentions, review volume, E-E-A-T signals. This is where traditional SEO skills apply, but only after layers one and two are sorted.

"Most businesses trying to improve AI visibility start at layer three and wonder why nothing changes. The answer is almost always that layer one or two has a problem that is blocking them."

Where to Start

Check layer one first: robots.txt allows AI crawlers, sitemap is current, pages load cleanly. Then layer two: Organisation and LocalBusiness schema on the homepage, Google Business Profile complete and accurate, consistent business data across directories. Only then does investing in content and link building make sense.

Common Questions About AI Search

How does AI search decide which businesses to recommend?
+
AI search platforms evaluate businesses against several criteria: structured data presence, content quality and relevance, off-site authority signals such as Reddit and directory mentions, and technical accessibility. Businesses that score well across all these dimensions are cited most often.
Does ChatGPT search the internet in real time?
+
ChatGPT with web browsing enabled does search in real time using Bing's index. For businesses, both your on-page content and your structured data need to be accessible to Bing and OAI-SearchBot for your business to appear in ChatGPT responses.
What is the difference between ChatGPT, Perplexity, and Google AI Overviews?
+
All three generate direct answers rather than lists of links, but work differently. ChatGPT combines training data with real-time Bing search. Perplexity is search-first with transparent citations. Google AI Overviews draws on the full Google index and Knowledge Graph, making schema markup and Google Business Profile particularly influential.
How do I get my business cited in AI search answers?
+
Add Organisation and LocalBusiness schema markup, ensure AI crawlers are allowed in robots.txt, complete your Google Business Profile, build verified external references, and create content answering your customers' questions. Most businesses see AI citation begin within four to eight weeks of implementing structured data.
Why does Google AI Overviews sometimes skip high-ranking pages?
+
AI Overviews prioritise pages that directly answer the query in structured, machine-readable format. A page can rank in position one on traditional Google while being skipped by AI Overviews if it buries the answer in unstructured prose or lacks schema markup. This is why Semrush found only 51% overlap between AI Mode citations and Google's top 10 results.