Why ChatGPT Doesn't Recommend You Yet (And How to Fix It)

If ChatGPT never names your business when someone asks for a recommendation in your category, you're not alone — and it's usually fixable. Answer engines don't have a "pay to play" switch for most local queries. They retrieve from the public web, then write. When you're missing from the answer, one or more of those steps is failing for you.

This is a practical how-to: diagnose the miss, fix the highest-leverage problems, and measure whether ChatGPT (and other answer engines) start mentioning you. The same checklist applies under the AEO / GEO label.

Step 1: Prove the problem with a question set

Don't rely on one prompt you typed once. Build a short list of real questions customers ask:

• "Best [service] near [city]"
• "Who should I hire for [specific job] in [city]?"
• "[Service] companies that [niche detail]"

Ask each question on ChatGPT with web browsing or search enabled (wording depends on the product version). Repeat a few days later. Record whether you appear, whether competitors appear, and which URLs get cited. Appearance rates under 50% for the same prompt across days are common — one local-recommendation half-life study found only about 48% of businesses resurfaced a few days later — so trends matter more than a single screenshot.

Step 2: Check whether ChatGPT can even see you

1. robots.txt: Make sure you aren't blocking GPTBot / OAI-SearchBot (or the current OpenAI crawler names in OpenAI's bot docs) unless you intend to. Many teams block training bots but still allow search crawlers — that's fine if intentional.
2. Indexing: Can a logged-out browser load your key service pages? Soft 404s, login walls and aggressive bot challenges all hide content from retrieval.
3. Homepage clarity: Does the first screen say who you are, where you work and what you do in plain language?

Step 3: Fix thin or contradictory pages

Engines prefer passages they can quote. Common misses:

• Service pages that are three vague paragraphs and a form
• City pages that don't mention the city until the footer
• Different phone numbers or hours on the site vs Google Business Profile
• Claims with no specifics (process, timeline, credentials)

Rewrite so the main question is answered in the first two sentences, then support it with headings, FAQs and concrete facts. Add schema that matches visible content — don't invent ratings.

Step 4: Look at who ChatGPT already cites

When ChatGPT answers your test questions, open the citations. You'll often see:

• Review sites and map listings
• Local news or chamber pages
• Trade directories and "best of" roundups
• Competitor sites with clearer comparison content

Your job is not to spam those sites. It's to earn accurate, up-to-date presence where the engine already looks — complete profiles, real reviews you ask for honestly, guest articles with facts, corrections when a directory has the wrong number.

Step 5: Measure AI referral traffic

In your analytics, create a segment or filter for referrals from domains associated with AI products (for example chat.openai.com, chatgpt.com, perplexity.ai, and similar). Numbers will be small at first. What you want is direction over months, not vanity spikes. Pair referral data with your manual appearance-rate table so you know whether "more traffic" matches "more mentions."

For a fuller definition of the practice behind this checklist, read what is answer engine optimization.

Want a structured baseline across ChatGPT and other answer engines? See the AEO audit or book a call.