How to get cited in ChatGPT as a local business.

Getting cited in ChatGPT as a local services business requires three things most local sites don't have: structured Q&A content that matches the shape of how people ask LLMs questions, schema markup that machines can parse without ambiguity, and a fresh signal layer (Google Business Profile activity, recent content, recent reviews) that tells LLMs your business is currently operating. Once those are in place, expect 60–90 days for first citations to appear.

What ChatGPT citation actually looks like

When a user asks ChatGPT "best plumber in Portland" or "HVAC company in Seattle for heat pump install," ChatGPT now returns a short answer that names three or four specific businesses. Each name is often clickable, often with the company URL. The three citations are picked from a much larger pool of candidates that ChatGPT considered — and most local services businesses aren't in the candidate pool at all because their content isn't structured the way LLMs need.

This is different from Google AI Overviews (which show up in Google search results) and different from Perplexity (which functions as a dedicated answer engine). The mechanisms overlap heavily — same underlying patterns help with all three — but ChatGPT specifically rewards a few patterns more than the others.

How LLMs decide who to cite

A useful mental model: LLMs are statistical pattern matchers that have been trained on a frozen snapshot of the web, then connected to live search at query time. When a query comes in, the LLM looks for content in the live web (and its training set) that's already shaped like an answer to that query. Pages that match the shape get pulled into the response. Pages that don't match the shape — even if they have the same information — get ignored.

The shape that matches looks like this: a clear question, a direct answer in 2–4 sentences, supporting structured detail, machine-readable metadata that confirms what the page is about. Local services pages that bury this shape under hero images, brand storytelling, and stock photography don't match — even if everything an LLM would need to cite is technically on the page.

Three patterns that get cited

1. Direct-answer paragraph at the top of every service page

Below the H1, before any image, write a 60–80 word paragraph that literally answers "what does this business do, where, and how." First sentence answers the question. Subsequent sentences add specificity. No marketing copy. No company history. No "Welcome to..."

This is the paragraph LLMs quote. It's also the paragraph that performs better in Google AI Overviews, in Perplexity, and in classic featured snippets. One change, three surfaces. The format is unusual enough on most local services sites that even a single page with this structure dramatically increases citation probability.

2. FAQ blocks with FAQPage schema

Pick the questions a buyer actually asks before contacting you. For local services, those questions cluster around cost, timing, scope, geography, financing, and credentials. Write 5–7 of them per major service page. Each answer 2–4 sentences. Wrap the entire block in FAQPage JSON-LD schema.

Schema is the part most local sites skip. Body copy with question-and-answer formatting helps a little. Body copy plus schema helps dramatically more. LLMs explicitly look for FAQPage schema as a signal of citable content because the schema confirms the question and answer are structurally what they appear to be — not just text that happens to read like a Q&A.

3. LocalBusiness + Service schema everywhere

LocalBusiness schema tells LLMs what kind of business you are, where you operate, and how to identify you. Most local sites add LocalBusiness only on the home page. Service schema describes the specific service the page is about. Most local sites don't add Service schema anywhere.

Adding both to every service page closes the gap LLMs need to specifically cite a page for a specific query. Without page-level Service schema, an LLM sees "a plumber in Portland" but doesn't know which specific service the page is about — so it might cite a generic competitor whose page is more clearly labeled. With Service schema, the page is unambiguously "emergency drain cleaning in Portland" and gets cited specifically for that query.

The freshness signal

ChatGPT and other LLMs combine training-data signals with live web search. The live signal includes things like Google Business Profile activity, recent reviews, recent content publishing dates, and crawl recency. A business that hasn't published anything in 8 months looks dormant to LLMs even if it's actively operating.

Three freshness habits that compound:

Weekly Google Business Profile posts. Cadence beats volume. One post per week consistently outperforms five posts in one day followed by silence.

Monthly blog content. Even a single 800–1500 word piece per month, written to answer a specific query, signals that the business is actively producing content. This piece you're reading is itself a freshness signal.

Recent reviews. Encourage one review per week from happy customers. The freshness of reviews matters as much as the count — a business with 200 reviews from 2022 looks worse than a business with 50 reviews dated within the last 60 days.

What doesn't work

A few approaches that local services advice from 2022 still recommends but don't help with LLM citation:

Stuffing service-area cities into a single page. "We serve Portland, Vancouver, Salem, Eugene, Bend, Tacoma, Seattle, Spokane, and surrounding areas..." LLMs don't cite pages that try to be about everywhere. They cite pages that are clearly about one place.

Mass directory submissions and citation services. They don't help with LLM citation. They probably stopped helping with classic local SEO years ago. If you're paying for a citation-submission service, cancel.

Long blog posts with no structure. A 2,500-word post that doesn't break into clear question-and-answer sections won't get cited because LLMs can't extract a clean quote from it. If you're writing long posts, structure them so each section has a clear question header and a quotable answer.

How to know if it's working

Three measurements, in order of usefulness:

Manual citation check. Once a week, ask ChatGPT (and Perplexity, and Google AI Overviews) the queries you want to be cited for. Note the three businesses cited. Track over time. This is the only direct read on LLM citation today — there's no analytics dashboard for it yet.

GBP "discovery search" growth. Google Search Console for GBP shows non-branded "discovery searches" — searches where someone found you without searching your name. Discovery search growth is a strong leading indicator that AI surfaces (Google AI Overview specifically) are starting to cite you, because most AI Overview clicks come through discovery rather than branded search.

Inbound calls from new geographic origins. If you start getting calls from cities you weren't getting calls from before, AI surfaces are likely the cause — they cite businesses across a broader geographic radius than classic Google search did, especially for "near me" queries.

FAQ

Is ChatGPT citation different from Google AI Overview?

The mechanisms overlap heavily — same underlying patterns help with both. ChatGPT specifically weights structured Q&A and FAQ schema heavily. Google AI Overviews weight a mix of schema, content quality, and traditional ranking signals. Perplexity weights citation density (how many other sources cite a page). Optimize for the patterns above and you get most of the benefit on all three.

How long does it take to get cited?

60–90 days from making the changes to first citations. Honest range. The mechanism is gradual — LLMs need to recrawl your pages, reindex them, and build statistical confidence that you're a citable source.

Can I just pay to get cited in ChatGPT?

No. There's no paid placement in ChatGPT citations. The mechanism is structural — pages that are shaped right get cited, pages that aren't don't, regardless of advertising spend.

Do reviews matter for ChatGPT citation?

Reviews matter as a freshness signal more than as a count signal. A business with 50 recent reviews looks better to LLMs than a business with 500 reviews from 3 years ago. Encourage steady review collection rather than a one-time push.

Will this work for my specific industry?

The patterns above are general — they apply to any local services business. The specific implementation varies by industry (HVAC FAQ schema looks different from plumbing FAQ schema looks different from electrical). The structural mechanism is the same.

Where to start this week

Pick one service page. Your highest-traffic one. Three changes:

1. Write a 60–80 word direct-answer paragraph at the top, below the H1, above any hero image.

2. Add 5 FAQ Q&A pairs at the bottom, wrapped in FAQPage JSON-LD schema.

3. Add LocalBusiness + Service schema to the page head if it's not already there.

Resubmit the URL in Google Search Console. Run a manual ChatGPT citation check on the related query today (baseline). Run it again in 30 days. Run it again in 60 days. Watch what changes.