Dank Rank Dispensary AEOStart with the audit
← Back to field notes

September 2, 2026

Why ChatGPT Is Ignoring Your Dispensary (And How to Fix It)

Why ChatGPT Is Ignoring Your Dispensary (And How to Fix It)

Clean dispensaryAEO-style interface graphic with icons for AI answers, local search, and dispensary entity signals

Your dispensary can rank in Google and still be absent from ChatGPT.

That is not a contradiction. It is the current search environment.

Google, ChatGPT, Perplexity, Gemini, and Google AI Overviews do not use the same sources, retrieve information in the same way, or present businesses through the same system. A strong Google Business Profile helps. It does not guarantee that an AI assistant can identify, verify, or cite your dispensary.

The visibility gap is measurable.

A 2026 SOCi Local Visibility Index analysis reviewed more than 350,000 business locations. ChatGPT recommended 1.2% of those locations. Perplexity recommended 7.4%. Gemini reached 11%. Google’s local 3-pack surfaced 35.9%.

AI does not always rank you lower.

Sometimes it leaves you out.

First, separate policy from visibility

ChatGPT may refuse a direct cannabis recommendation even when your dispensary has a clean digital footprint.

OpenAI’s commerce policies prohibit commerce features from listing or promoting cannabis and THC products. Its ad policies also restrict advertising for cannabis and other intoxicating substances.

That matters. A direct prompt such as “Which dispensary should I buy cannabis from?” can produce a refusal, a neutral answer, or a referral to a directory. No amount of schema markup can force an assistant to make a recommendation it will not provide.

But policy is not the whole explanation.

ChatGPT and other assistants also answer many adjacent questions:

Your store can be named, cited, or included in these answers only if the assistant has enough reliable information to connect the query to your business.

That is the work.

Google visibility is not assistant visibility

Your Google Business Profile is one surface. It is not the entire entity.

Different assistants use different combinations of:

The practical consequence is simple:

A Google win does not transfer automatically to ChatGPT.

A dispensary may have:

…and still have an outdated address on another listing, no location-specific schema, or no clear connection between its store page and parent brand.

AI systems do not just ask, “Does this page rank?”

They also ask, “Is this the same business? Is it licensed? Is the information current? Can I verify the answer from more than one source?”

That is entity confidence.

Abstract AI search interface showing query results, rankings, and citation status across assistant columns

Fix 1: Run the queries shoppers actually ask

Do not begin with a generic SEO audit.

Begin with the answer.

Ask ChatGPT, Perplexity, Gemini, and Google AI Overviews the same local and product-intent questions your customers ask. Record the response exactly.

Examples:

Then record:

Field What to capture
Assistant ChatGPT, Perplexity, Gemini, or Google AI Overviews
Query The exact prompt asked
Named stores Every dispensary included
Store framing How each business is described
Citation URL The source used in the answer
Your status Named, cited, mentioned, or absent
Competitor pattern Which sources appear repeatedly

Run each query more than once. AI answers can change.

The goal is not a screenshot collection. The goal is a query record that shows where you are named, where you are missing, and which competitor sources keep appearing.

We call this a query sweep.

Fix 2: Close the citation gap

If a competitor appears in the answer and you do not, inspect the citations.

A citation gap is not simply “your competitor has better SEO.” It is the difference between the sources supporting their entity and the sources supporting yours.

Review:

Compare the basics:

One old address can create doubt. One duplicate listing can split authority. One stale menu can make a current product claim impossible to verify.

Do not treat citations as a list of links you collected years ago.

Treat them as active business records.

Vector illustration of a dispensary entity connected to business listings, directories, maps, and editorial sources with missing citation links

Fix 3: Make your dispensary one clear entity

Your website should make it easy to answer 5 questions:

  1. What is the legal or public-facing business name?
  2. Where is the location?
  3. When is it open?
  4. What does it offer?
  5. Which pages and listings confirm the same information?

Each location needs its own crawlable page. Do not hide every store behind a JavaScript locator or a single generic “locations” page.

A location page should include:

For multiple locations, connect each store to the parent brand without blending the locations together. Your Detroit store, Grand Rapids store, and Ann Arbor store are related entities. They are not one address.

This distinction matters for multi-location operators. A single brand page cannot replace clean location records.

Fix 4: Replace the PDF menu with structured data

Most operators publish PDFs.

We publish data.

A PDF may be useful for a customer who wants to download a menu. It is a weak primary data source for an assistant trying to understand products, categories, availability, and location.

Use crawlable HTML for core menu information. Then support it with accurate schema markup.

Google’s official LocalBusiness structured data documentation recommends marking up details such as the business name, address, phone number, hours, URL, and other location information. Google also says that AI Overviews do not require a special AI schema. The fundamentals still matter: crawlable pages, visible text, accurate structured data, and current Business Profile information. See Google’s AI features documentation.

For a dispensary, the implementation should be reviewed across:

Schema is not a shortcut.

It is a machine-readable version of information already present on the page. If your markup says one thing and the visible page says another, you have created another confidence problem.

Structured menu data interface transforming an unstructured document into organized product, offer, and location data cards

Fix 5: Keep your Google Business Profile current

Google still matters.

Google’s own guidance recommends keeping Business Profile information up to date and making important content available in text. AI features may use query fan-out, meaning the system runs several related searches across different sources before forming an answer.

Your profile is part of that ecosystem.

Check:

Do not stop after the initial setup.

A stale profile is not a completed profile. Review it monthly. Record changes. Update the website and third-party listings at the same time.

Fix 6: Build proof outside your own website

Your website explains who you are.

Third-party sources confirm it.

AI assistants need external references when they assemble local answers. Build a footprint that includes legitimate, relevant mentions:

Do not buy a package of generic directory links. Do not publish unsupported medical claims. Do not create thin city pages with the same paragraph copied 20 times.

Build records a shopper and an assistant can verify.

What to do this week

Use this order:

  1. Run 20–40 local and product-intent queries across 4 assistants.
  2. Record every named dispensary and every citation URL.
  3. Compare your business details across 8–12 relevant listing surfaces.
  4. Claim and correct missing or outdated profiles.
  5. Create or improve each location page.
  6. Replace PDF-only menu information with crawlable HTML.
  7. Add and validate LocalBusiness, Organization, Product, and menu-related structured data.
  8. Set a monthly review for citations, hours, menus, and entity connections.
  9. Re-run the same query set after fixes ship.
  10. Keep the report.

Do not guess whether ChatGPT is ignoring you.

Query it.

Start with a written citation-gap report

You do not need a 12-month contract to find the problem.

The dispensaryAEO AEO audit is a one-time $500 engagement with an approximately 2-week turnaround. We query ChatGPT, Perplexity, Gemini, and Google AI Overviews using Michigan shopper prompts.

You receive:

You keep the report whether you continue with us or not.

The person who reads your submission is the person who runs the queries, writes the findings, and answers the next email. Submit the audit intake. We reply within 1 business day.

Your dispensary does not need to be named in every answer.

It does need to be verifiable when the assistant can safely answer.

That starts with data, citations, and a report; not a guess.

Sources