How to Get Cited in ChatGPT: The 6 Citation Signals and a 30-Day Action Plan

Why some businesses get cited and most don’t. ChatGPT does not cite randomly. It cites sources that are authoritative, structured, consistent, and specific. Most businesses fail on at least three of those four. The businesses that appear regularly in ChatGPT answers have built six citation signals that most of their competitors have not touched. This guide explains each signal and gives you a 30-day plan to build them.

We run AI visibility audits for local businesses and professional services. The question we get asked most often is some version of “why does my competitor show up in ChatGPT and I don’t?” The answer is almost always the same six signals, and most of them can be improved in under 30 days.

---

How ChatGPT cites sources

Before getting into the tactics, it helps to understand the two modes ChatGPT operates in, because they affect different citation paths.

Mode 1: Training data (standard ChatGPT)

Standard ChatGPT draws from a training corpus assembled from web content before the model’s knowledge cutoff. When ChatGPT answers a question without web browsing enabled, it is drawing on patterns and facts absorbed during training. It may mention specific businesses or websites that appeared frequently and authoritatively in that training corpus. It does not always cite sources explicitly in this mode.

Mode 2: Web browsing (ChatGPT Plus with Browse)

When web browsing is enabled, ChatGPT runs live queries against Bing’s search index, reads the top results, and synthesises an answer with explicit citations. This is similar to how Perplexity works. The difference from Google AIO is that ChatGPT uses Bing, not Google.

For most businesses trying to get cited, Mode 2 is the more tractable target. You can improve your Bing indexing and content structure faster than you can change what is in OpenAI’s training corpus. The citation improvements you make now will also carry into future training cycles as those pages get crawled and indexed.

What triggers ChatGPT to include a citation

In browsing mode, ChatGPT cites sources when:

• The source directly answers the query
• The source is clearly authoritative (high-authority domain, specific expert, verified business)
• The source contains structured, extractable information (rather than vague descriptive copy)
• Multiple queries lead to the same source (signals high relevance)

In training mode, a business becomes part of ChatGPT’s implicit knowledge when:

• It is mentioned frequently in high-authority web content (news, directories, review platforms)
• It is associated with specific geographic locations and service categories consistently
• Its own website has clear, authoritative answers to category questions

---

The 6 citation signals

Signal 1: Entity authority

Entity authority is the most important signal and the hardest to build quickly. It is the sum of how consistently and authoritatively your business is represented across the web.

ChatGPT learns about businesses from the aggregate of web content it has seen. A business that is mentioned in local news, featured in professional association directories, reviewed on multiple platforms, and cited by other businesses has higher entity authority than an identical business with only a website and a Google Maps listing.

What builds entity authority:

• Google Business Profile - complete, accurate, regularly updated
• Major directory listings (Yelp, BBB, HomeAdvisor, industry-specific directories)
• Local news coverage - even a single mention in a local newspaper carries meaningful weight
• Professional association membership pages (if you are a member, your profile on their site is a citation asset)
• Review platform profiles with substantial review volume
• LinkedIn company page with complete information
• Wikipedia or industry wiki mentions (the highest-value citation source for training data)

How to check your entity authority: Search ChatGPT for “[your business name] [your city]”. Does it know you exist? What does it say? If it either draws a blank or confuses you with another business, your entity authority is low.

Signal 2: Structured data

Schema.org markup is the clearest signal you can give to a crawling system about what your content contains and how it is structured. Both PerplexityBot and Bingbot (which feeds ChatGPT’s browsing mode) parse schema markup when indexing pages.

The schema types that most directly improve ChatGPT citation probability:

FAQPage: The single highest-return schema type. Mark up every Q&A section with FAQPage schema and ChatGPT can extract your questions and answers directly as citation candidates.

LocalBusiness: Declares your business name, address, phone, service area, and category to every system that reads it. Without this, ChatGPT is guessing at your location and service scope.

Service: For each specific service you offer, a Service schema block declares what it is, who provides it, and where. This helps ChatGPT associate your business with specific service category queries.

HowTo: For guides, walkthroughs, and process pages. ChatGPT particularly values HowTo schema because the steps structure maps cleanly to answer generation.

Person: For professional services - lawyers, dentists, financial advisors - Person schema with credentials and affiliation helps establish the human expertise behind the business.

The most common schema gap we find in audits: sites have Article or Organisation schema but are missing FAQPage, Service, and LocalBusiness on the pages that would benefit most. Adding these three types to your top 5 pages is a 1-2 day implementation task.

Signal 3: Third-party mentions

ChatGPT is trained on web content. The more often your business name, website, and expertise are mentioned in content that appears across the web, the more likely ChatGPT is to include you in relevant answers.

This is analogous to backlinks for traditional SEO, but the mechanism is different. With backlinks, link equity flows through the link graph and boosts your domain’s PageRank. With third-party mentions for ChatGPT, frequency and quality of mentions in authoritative text teaches the model to associate your business with specific topics, locations, and expertise.

The mention sources that matter most:

• Local news articles - a mention in a regional newspaper carries more training data weight than 50 directory listings
• Industry publications - a featured case study or contributed article in a trade publication establishes category authority
• Professional association pages - membership directories and featured member profiles
• High-authority review aggregators - Houzz, Angi, Thumbtack, and category-specific platforms
• University or government resources - if your business or expertise is referenced in an educational or governmental context, this is extremely high-value
• Other business websites - suppliers, partners, and clients that mention you on their sites

What not to chase: Low-authority directory spam. A business with 200 directory listings on aggregator sites that themselves have no authority does not build meaningful entity presence. Quality and authority of the mention source matter more than volume.

Signal 4: Review volume and specificity

Review signals influence ChatGPT citations in two ways. First, review volume and quality are part of the entity authority picture - a business with 500 reviews at 4.8 stars is clearly a real, active, trusted operation. Second, the text content of reviews adds to the web representation of your business. Reviews that mention specific services, locations, techniques, and outcomes are richer training data signals than generic “great service” reviews.

What ChatGPT-friendly reviews look like:

• Mention the specific service (“They replaced my entire HVAC system in one day”)
• Reference the location (“The technician drove from Denver to our home in Arvada”)
• Include a specific outcome (“Our energy bills dropped 30% after the insulation job”)
• Name the staff member where appropriate (“Ask for Marcus - he is extremely thorough”)

You cannot write reviews for yourself. But you can ask satisfied clients to describe their specific experience rather than leaving a generic rating. That specificity feeds into both Google’s quality assessment and, over time, ChatGPT’s understanding of your business.

Review platform targets:

• Google Business Profile (highest weight for local AI citations)
• Yelp (significant for consumer services)
• Industry-specific platforms (Houzz for contractors, Avvo for legal, Healthgrades for medical)
• Facebook (lower weight but contributes to aggregate entity presence)

Signal 5: Geographic signals

Geographic specificity is a signal that is often missing from local business content. ChatGPT needs to know where you operate, what areas you serve, and what your physical location is. Vague geographic claims (“serving the greater metropolitan area”) are harder for AI systems to parse than specific declarations (“serving Dallas, Plano, Frisco, and McKinney within a 30-mile radius”).

Where to declare geographic signals:

• areaServed field in LocalBusiness schema
• Service area paragraphs on the homepage and service pages
• Location-specific content pages for each major service area city
• Google Business Profile service area settings
• Bing Places service area configuration
• Directory listings with consistent service area declarations

Location page content strategy: If you serve 10 cities, having a dedicated page for each city is not just an SEO play - it is a GEO signal. Each page establishes the association between your business and that location in the content index. ChatGPT can cite “Best movers in Plano” with your Plano page as the source.

Signal 6: Content directness

The content pattern that ChatGPT cites most reliably: a page that begins with a 40-80 word paragraph that directly answers a specific question, followed by structured sections that expand on that answer.

The academic research backing this up is the arXiv GEO paper (Aggarwal et al., 2023), which tested 10 content intervention strategies across multiple LLM systems. “Fluency” interventions - rewriting content to be clearer and more direct - consistently increased citation frequency. “Statistics” interventions - adding specific data points - also increased citation rates significantly.

What directness means in practice:

• Open with the answer, not the context
• Use specific numbers (prices, timeframes, distances, percentages) rather than vague qualifiers (“affordable”, “quick”, “local”)
• Write sentences that contain one idea each
• Use H2 and H3 headers that describe what is in each section (not clever marketing headers)
• Include a FAQ section at the bottom of every substantive page

The comparison that illustrates this:

Not citable:

“Finding the right moving company can be a stressful process. At City Movers, we understand how important your belongings are to you, which is why we offer comprehensive moving services designed to make your relocation as smooth as possible.”

Citable:

“Local moving companies in Chicago typically charge $95-$145/hour for a 2-person crew. A standard 2-bedroom move within the city takes 4-6 hours. Long-distance moves from Chicago to other Illinois cities average $800-$1,600 depending on distance and volume. City Movers serves all Chicago neighborhoods plus Evanston, Oak Park, and Naperville.”

The first version describes a business’s philosophy. The second answers a question. ChatGPT cites the second type.

---

How to run a citation audit

A ChatGPT citation audit takes about 2 hours and costs nothing if you have ChatGPT Plus (for browsing mode) or access to a free account you can use for manual testing.

Build your query set

Write 30 queries across three categories:

Category queries (10 queries): “best [service] in [city]” “top-rated [service] [city]” “[service] cost [city]” “how much does [service] cost [city]” “[service] near me [city]” (substitute your city)

Problem queries (10 queries): “[specific problem] [city]” “how to [task your clients need]” “[emergency/urgent service] [city]” “[comparison] [city]” (e.g. “move or stay [city]”) “what to look for in [service provider]”

Informational queries (10 queries): “what is [core service]” “how does [process] work” “[industry term] explained” “[service] checklist” “[service] questions to ask”

Run and record

For each query:

1. Run in ChatGPT with browsing enabled (Plus account)
2. Note whether your domain is cited
3. Note which domains are cited if not yours
4. Copy the exact answer text for queries where you are absent - this tells you what ChatGPT currently knows about your topic

Calculate your citation rate

Count your citations across all 30 queries. Divide by 30. That is your current ChatGPT citation rate.

Benchmarks:

• Under 5%: baseline, no GEO work done
• 5-15%: some foundational signals present, room for significant improvement
• 15-30%: solid work done, competitive in most local markets
• 30%+: strong GEO presence, focus on expanding to more query types

Identify gaps

For every query where you are absent, note the cited source. Visit that page and assess: what does it do differently? Specifically:

• Is the opening paragraph a direct answer?
• Does it have FAQPage schema?
• Does it have specific data points?
• Is the author or business named with credentials?

That comparison gives you the exact gap to close.

---

The 30-day action plan

This is the sprint plan we run for new clients who want to improve ChatGPT citation rate quickly. It is structured so the highest-leverage work comes first.

Days 1-3: Baseline audit

Run your 30-query citation audit and record your baseline. Set up a simple spreadsheet with query, current cited domain (if not yours), and notes on what the cited page does well. This becomes your progress-tracking tool for the whole 30 days.

Days 4-10: Fix entity signals

Google Business Profile:

• Verify all information is current and accurate
• Add all services you offer
• Upload 10+ recent photos
• Confirm service area is correctly set

Bing Places:

• Create or claim and verify your Bing Places listing
• Confirm it matches your Google Business Profile exactly

NAP consistency audit:

• Check your business name, address, and phone on Google, Yelp, BBB, HomeAdvisor, and 10 other directories
• Correct any inconsistencies (this is the most boring and most impactful work on this list)

Days 7-14: Add structured data

Pick your top 5 pages (homepage, top 3 service pages, and one content page).

For each page, add:

• LocalBusiness schema (homepage only - includes full address, phone, areaServed)
• Service schema (each service page)
• FAQPage schema (any page with Q&A content)

Validate each schema block in Google’s Rich Results Test. Fix any errors before moving on.

Days 10-20: Rewrite content as direct answers

Take your 5 priority pages and rewrite the opening paragraph of each one as a direct-answer capsule. The capsule should:

• Be 40-80 words
• Answer the primary question the page targets
• Include at least 2 specific data points (prices, timeframes, measurements, or percentages)
• Mention the city or service area
• Not start with “We” or a description of the business

Then add a FAQ section at the bottom of each page with at least 5 questions. Write these as real questions your clients ask, with direct answers. Mark them up with FAQPage schema.

Days 15-30: Build third-party citations

Week 3:

• Submit to 10 new high-authority directories you are not currently listed in. Focus on industry-specific directories (Houzz for contractors, Healthgrades for medical, etc.) and local business directories (chamber of commerce, local newspaper business directory).
• Draft a short press release about a recent project, award, or milestone. Distribute to your local newspaper’s business desk and one industry publication.

Week 4:

• Contact one industry association you are eligible to join (or already belong to). Ensure your listing on their member directory is complete.
• Reach out to 3 businesses you have partnered with or done work for and ask for a case study mention or testimonial on their website.

Day 30: Re-audit and set next sprint

Re-run your 30-query citation audit. Compare to baseline. Most clients see 5-15 percentage point improvement after this sprint - not a sweep, but a meaningful baseline shift.

For the queries still showing gaps, set the next sprint priorities. The work after Day 30 is usually: expand to more location pages, build more third-party citations, and add more topic-specific content for the query categories where you are still absent.

---

FAQ

Can any website get cited in ChatGPT?

Yes. ChatGPT can cite any publicly indexed website when it uses web browsing mode. In non-browsing mode, ChatGPT draws from its training data, which includes content indexed before its knowledge cutoff. Getting cited regularly requires building the signals that make your content authoritative, structured, and extraction-friendly.

Does ChatGPT use Google to find sources?

ChatGPT with browsing enabled uses Bing’s search index, not Google’s. In standard mode, it draws from its training data corpus. Building presence in Bing’s index and high-authority training data sources (Wikipedia, major publications, industry directories) maximises ChatGPT citation probability.

Why does ChatGPT cite some businesses but not others?

ChatGPT is more likely to cite businesses with strong entity authority (consistent mentions across authoritative sources), well-structured direct-answer content, high review volume and quality, geographic specificity in their content, and clear structured data.

Does schema markup help with ChatGPT citations?

Yes. Schema markup - especially FAQPage, LocalBusiness, Service, and HowTo - helps ChatGPT extract structured answers more accurately when in browsing mode. In training mode, schema-rich pages are more reliably parsed and attributed during corpus assembly.

How long does it take to appear in ChatGPT answers?

In browsing mode, improvements in Bing indexing and content structure can show results within 30-60 days. In standard training-data mode, changes propagate on OpenAI’s retraining schedule (historically every 6-12 months). Targeting browsing-mode optimisation is the faster path.

---

Related guides

• How to rank in Perplexity
• Google AI Overviews for local businesses
• What is Generative Engine Optimisation (GEO)?
• Local SEO in the age of AI search

---

Published by Vespio. We run AI visibility audits and citation-building sprints for local businesses. If you want to know exactly what ChatGPT says about your business today, ask us for a diagnostic.