How It Works
The methodology behind every Clever Merchants citation
AI citation is not random. It follows from five specific layers of work. This page explains exactly what each layer is, why it matters, and how they combine to produce verifiable citations in ChatGPT, Perplexity, and Google AI Overviews.
What generative engine optimisation actually is
When a buyer asks ChatGPT "who is the best commercial electrician in Dublin?" or "which software agencies are worth talking to?", the AI is not running a search. It is generating an answer based on patterns in its training data — web pages, directories, schema markup, review sites, publications — weighted by how consistently and confidently those sources describe a given business.
Traditional SEO optimises for Google's ranking algorithm: backlinks, domain authority, keyword density, page speed. Those signals matter for search engine results pages. They have almost no bearing on whether an AI tool cites your business.
Generative engine optimisation addresses the signals AI systems actually use: structured machine-readable data about your business, content that answers buyer questions directly, a consistent verified identity across multiple external sources, and proof that you are an established credible operator in your category.
The good news is that these signals are implementable. They are not expensive relative to the value of being cited, and they do not conflict with traditional SEO. The Clever Merchants methodology delivers all five in a single programme.
The five implementation layers
Each layer addresses a different signal. Together they create the conditions for consistent AI citation.
Schema markup — the technical foundation
Schema markup is machine-readable code added to your website that tells AI crawlers exactly what your business is, what it offers, who it serves, and where it operates. AI systems read structured data before body copy. A business with no schema is harder to classify and cite confidently.
Schema types deployed in the Clever Merchants programme:
Organization
Business identity, contact details, logo, founding date — the baseline entity record that all other schema references
Service / ProfessionalService
What you do, for whom, in which geographic area. Each service gets its own Service schema node
FAQPage
Every FAQ block on the site. AI systems pull verbatim answers from FAQPage schema when answering buyer questions
LocalBusiness
For businesses serving a defined area. Feeds into Google AI Overviews and local AI query answers
Person
Founder and key personnel — establishes the human expertise behind the business
Article / BlogPosting
On all published blog posts and guides. Signals that the content is authored, dated, and attributed
BreadcrumbList
Site structure — helps AI systems understand the relationship between pages
Product (ecommerce)
Product name, description, price, availability on collection and product pages
Question-answering content
AI systems are trained to answer questions. They cite businesses whose content is structured around the questions buyers actually ask. Most business websites are written as brochures. GEO-optimised pages are written as citable sources.
Content formats that produce citations:
FAQ sections on service and product pages
5–8 questions per page, each answered directly in 2–4 sentences. Marked up with FAQPage schema. These are the most reliably cited content format across all AI tools.
Service page rewrites
Pages rewritten to answer "what do you do, for whom, why are you qualified, what outcomes can a buyer expect" — the four questions AI tools are answering when they recommend a business.
Buying guides
Long-form content that answers comparison and evaluation questions. AI systems most frequently quote from buying guides verbatim because they contain the most specific, actionable answers.
Authority blog posts
Definition pieces, how-to-choose pieces, and original research. Builds the citation footprint and gives AI tools content to attribute to your business.
Entity recognition — NAP consistency
AI systems work with named entities: things they can identify and describe with confidence. Your business is an entity. For an AI to cite it confidently, that entity needs to be described consistently across every source that mentions it.
What entity consistency requires:
Your Name, Address, Phone number (NAP) must be identical across every source: your website, Google Business Profile, Bing Places, Yelp, LinkedIn, Facebook, and every directory listing.
Even minor variations create conflicting signals. "Acme Electrical Ltd" on your website and "Acme Electrical" on a directory listing register as two different entities. AI systems treat conflicting identity signals as uncertainty — and uncertain businesses don't get cited.
The NAP audit is the least glamorous part of GEO and one of the highest-impact. It typically surfaces 6–12 inconsistencies even on well-maintained business profiles.
Citation footprint — external mentions
A business that exists only on its own website is, from an AI's perspective, unverified. The more credible third-party sources that mention your business — with consistent identity data and relevant context — the more confident an AI system is that you are established and worth recommending.
Citation footprint sources, by priority:
Tier 1
Google Business Profile, Bing Places, Apple Maps — feed directly into AI Overviews and local query answers
Tier 2
National and industry directories: Yelp, Trustpilot, Clutch (agencies), G2 (software), trade body listings
Tier 3
Local business directories, Chamber of Commerce listings, regional business databases
Tier 4
Press mentions, guest articles, podcast appearances — high-value but optional in the initial programme
E-E-A-T signals — expertise and trust
For high-stakes buyer queries — "which accountant should I use?", "who is the best solar installer in Cork?" — AI tools weight signals of genuine expertise and proven track record. A business with credentials, case studies, and a named founder is more likely to be cited than one that simply exists.
Experience
Years in operation, volume of work completed, named client outcomes — demonstrates real-world track record
Expertise
Founder credentials, certifications, accreditations, industry body memberships — structured in Person and Organization schema
Authoritativeness
Case studies with named outcomes, original research, quotes and mentions in industry publications
Trust
Consistent identity data, transparent contact information, real team members, published terms and policies
Why all five layers matter
Each layer addresses a different failure mode. A business with perfect schema but no FAQ content gives AI tools a machine-readable identity but nothing to quote. A business with excellent content but inconsistent NAP data confuses AI systems about which entity the content belongs to. A business with great content and schema but no external citations exists only in its own ecosystem — an AI tool has no third-party confirmation that it is real and established.
The combination is what produces confident citations. When an AI tool encounters a buyer query, it is looking for a business it can identify clearly (schema + entity recognition), describe accurately (question-answering content), and verify from multiple sources (citation footprint + E-E-A-T). All five layers need to be in place.
This is why partial GEO implementations often underperform. Content alone, schema alone, or directory listings alone are each insufficient. The Shamrock Electrical result — ChatGPT citation within 72 hours — came from implementing all five layers in one round of work.
How citations are verified
Every Clever Merchants programme includes a baseline citation audit before work begins and a final verification report when implementation is complete. Both use the same method: running the specific buyer queries relevant to your business through ChatGPT, Perplexity, and Google AI Overviews and recording whether your business appears.
The verification is self-evident. You can run the same queries yourself at any time. There is no proprietary metric, no black-box scoring — just the actual answer an AI tool gives when a buyer asks about your category.
The 90-day citation guarantee requires a minimum of three unprompted citations across ChatGPT, Perplexity, and Google AI Overviews on queries relevant to your business. If they are not present within 90 days of implementation, we continue working at no additional cost until they are.