ChatGPT does not rank businesses the way Google does. It recommends businesses it can identify, verify, and extract from — in that order. Understanding how each of those three things works tells you exactly what to fix.
"Ranking" in ChatGPT is the wrong frame. ChatGPT does not produce a ranked list the way Google does. It produces a recommendation — naming specific businesses — based on a decision process that has nothing to do with keyword position.
Think of it like the difference between the Yellow Pages and asking a knowledgeable friend. The Yellow Pages ranks businesses by how much they paid and how well they matched a search term. A knowledgeable friend recommends based on whether they know the business, whether they trust it, and whether they can explain why it is a good fit.
ChatGPT is the knowledgeable friend. The three things it needs — to know you, to trust you, and to have something specific to say about you — are the three signals this article covers.
Why does ChatGPT recommend some businesses but not others?
ChatGPT recommends businesses it is confident about. Confidence comes from three things working together:
- It knows who you are and what you do
- Other sources confirm you exist and are credible
- Your content gives it something specific to say when it recommends you
A business that fails the first test will not appear in recommendations — full stop. ChatGPT will not recommend a business it cannot clearly identify. A business that passes the first test but has no outside confirmation may get mentioned as a possibility but will not receive the specific, endorsing language that constitutes a real recommendation. A business that passes both but has no extractable content may get a recommendation, but the tool has to be vague about why — which produces a weaker result.
The good news: each of these is fixable with specific, practical steps.
Signal 1: Does ChatGPT know clearly who you are?
ChatGPT builds its picture of your business from the first content it can reliably associate with you. For most businesses, that is your website's homepage.
The main headline at the top of your homepage is the primary identity signal. If it says "Welcome to Smith & Co" or "Your trusted local experts," ChatGPT has no usable information. It knows a business called Smith & Co exists, but it cannot tell what the business does, who it serves, or where it operates. Without those details, it cannot match you to the queries your customers type.
A homepage headline that answers four questions in plain language gives ChatGPT the information it needs:
- Who you are
- What you do specifically
- Who you serve
- Where you operate (if location matters)
"Smith & Co: employment lawyers in Melbourne specialising in unfair dismissal claims for small business owners" is clear identity. "Trusted legal experts committed to your success" is not.
This is not about cramming keywords into your headline. It is about being specific. A specific identity gives ChatGPT something to match. A vague one gives it nothing.
Background code matters here too. Schema markup (say it: "SKEE-muh") is code in your website's background that declares your business identity to AI tools in machine-readable language. Without it, ChatGPT infers your identity from your visible content and tends to be conservative about businesses it cannot confirm. With it, you give the tool a verified anchor.
Any change that affects how ChatGPT reads your identity — a website redesign, a rebrand, a change in your service focus — requires rechecking the identity layer. The identity the tool has on record needs to match the business you currently are.
Signal 2: Do outside sources confirm you are real and credible?
ChatGPT does not recommend businesses based solely on what those businesses say about themselves. It looks for outside confirmation. It needs to know that the real world — not just your own website — backs you up.
Think of it like a reference check. Your resume says you are great at your job. But a hiring manager calls your references to confirm it. ChatGPT does the same thing, checking multiple outside sources before it will confidently endorse you.
Three categories of outside confirmation carry the most weight.
Google reviews. A pattern of recent, specific reviews from real customers on Google Business Profile functions as strong confirmation for service businesses. What matters is not just volume — it is recency and specificity. A review that names the specific service you provided and the outcome it produced is more confirming than ten generic "great service!" reviews. If your review profile is stale, the most direct fix is to ask current and recent customers to leave an honest review about a specific experience.
Mentions on credible outside websites. A listing in a relevant industry directory, a mention in a local news article, a profile on a professional association's website — these all confirm that you exist in the world, not just on your own site. Three to five credible, relevant mentions move the confirmation signal more reliably than dozens of generic directory submissions. Quality over quantity, always.
Professional association and certification listings. For regulated industries — law, accounting, finance, healthcare — being listed on an industry body's member directory is a particularly powerful form of confirmation. It is a categorical third-party verification that ChatGPT can read and act on.
The confirmation layer takes the most time to build. Reviews accumulate from genuine customer activity. Outside mentions come from consistent presence in your industry and community. There is no shortcut — which is exactly why businesses that start building this layer early accumulate a compounding advantage over time.
Signal 3: Does your content give ChatGPT something specific to say?
When ChatGPT recommends your business, it is more useful to the person asking if it can explain why. That explanation comes from something it has extracted from your content.
A business with no extractable content might still get recommended — but the tool has to say something vague like "they are a well-regarded option in this field." A business with rich, specific, extractable content gets a recommendation like "Smith & Co specialise in unfair dismissal claims for small businesses in Victoria and typically resolve cases within 60 days of lodging with the Fair Work Commission." The second recommendation drives action. The first does not.
Three content patterns produce reliable extractability.
Question-led headings with answer-first paragraphs. A page with headings phrased as questions — "What does an employment lawyer do for small businesses?" — followed by a paragraph that answers the question in the first sentence is structurally identical to what ChatGPT produces in its responses. The tool lifts this format readily. A page with topic-label headings — "Our Employment Law Services" — and paragraphs that open with background before the answer is harder to extract from and tends to be skipped.
FAQ sections with customer-language questions. A dedicated FAQ section is the single highest-yield content investment for ChatGPT extractability. The questions must be phrased the way a potential buyer would type them — specific, outcome-focused, and free of industry terminology. "What is the process for an unfair dismissal claim in Victoria?" is extractable. "What legal services do you provide?" is not.
Specific, named detail. ChatGPT extracts specific information — named processes, specific timeframes, explicit outcomes — far more readily than general claims. "Our team of qualified lawyers" is not extractable. "Employment lawyers typically resolve unfair dismissal claims within 60 days of lodging with the Fair Work Commission" is. The specificity gives the tool something concrete to cite.
An llms.txt file (say it: "L-L-M-S dot text") at your website root — yoursite.com/llms.txt — helps ChatGPT identify which pages to extract from and how to weight them. For sites with a mix of content types, it tells the tool which pages are authoritative for recommendation purposes. The format guide is at llmstxt.org.
What is the difference between being mentioned and being recommended?
There is a meaningful difference between appearing in a ChatGPT response and being recommended in one.
Mentioned means ChatGPT acknowledges your business exists. It may list you among several options, or include you in a general overview. The user sees your name. They do not receive an endorsement.
Recommended means ChatGPT names you with specific, positive context. "Smith & Co is a strong option — they specialise in unfair dismissal claims for small businesses in Melbourne and have a strong track record with Fair Work Commission cases." The user receives a specific reason to choose you.
Clear identity and background code tend to produce mentions. Outside confirmation and extractable content lift mentions to recommendations.
This distinction matters because customers act on recommendations, not mentions. A mention in a five-option list requires the customer to evaluate further. A recommendation reduces the decision for them. Getting from mentioned to recommended is the goal.
How do you test where you currently stand?
The right test uses need-based queries, not your own business name. Typing your business name into ChatGPT tells you whether the tool recognises you — not whether it would recommend you to a stranger.
The test that matters: "Who should I use for [your service] in [your location] if I am [your target customer type]?"
Run the same query across ChatGPT, Perplexity, Gemini, Claude, Grok and DeepSeek. Inconsistent results — recommended on one tool, missing on another — point to tool-specific gaps rather than fundamental problems. Missing across all tools points to the identity or confirmation layer.
The article on how to check your AI visibility walks through the full testing process step by step. Start there. Know where you stand. Then act on the signal that needs work.
Frequently asked questions
Is getting into ChatGPT answers the same as SEO?
No. Traditional search optimisation helps you rank on a results page so users can choose to click through to you. Getting into ChatGPT answers means being named in a single AI-generated recommendation. The two overlap on basic website quality but diverge significantly on what drives results. Google rewards keyword relevance and how many other sites link to you. ChatGPT rewards clear identity, outside confirmation and content it can extract from. Both are worth pursuing. They are different disciplines.
How long does it take to appear in ChatGPT recommendations after making changes?
Changes to your homepage headline and background code typically surface in ChatGPT responses within two to six weeks. For tools that check live websites — like Perplexity — changes can appear in days. Building trust signals takes longer because those depend on third-party actions. Most businesses that fix the top two or three issues see a real improvement within 30 to 60 days.
Does ChatGPT treat all business categories the same?
No. ChatGPT's willingness to recommend specific businesses varies by category. Service businesses with a clear local or specialist identity are recommended more readily than businesses in categories where ChatGPT is more conservative. Running the test queries described in this article tells you what ChatGPT is willing to produce for your specific field.
Can I pay to appear in ChatGPT recommendations?
No. ChatGPT does not sell paid placements inside its recommendation answers. Visibility is earned through site structure, outside confirmation, and useful content — not purchased.
What is the difference between being mentioned and being recommended in ChatGPT?
Being mentioned means your business name appears in a response without an endorsement. Being recommended means ChatGPT names you with specific positive context. Clear identity and background code tend to produce mentions. Outside confirmation and extractable content lift mentions to recommendations. Customers act on recommendations, not mentions.
