Question 1

Why won't ChatGPT recommend a cannabis dispensary by name?

Accepted Answer

ChatGPT, Claude, Gemini, and Perplexity all apply a content policy layer above the model that refuses direct retail recommendations for federally Schedule I substances, which still includes cannabis under United States federal law as of May 2026 despite the pending HHS recommendation to reschedule to Schedule III. The refusal is consistent across all 24 recreational and 38 medical states because the policy enforcement happens at the global filter level, not at the user-location level. When a user asks the model to recommend a dispensary in Denver or Tampa or Las Vegas, the system prompt instructs the model to decline, suggest the user check Leafly or Weedmaps, and offer educational information about strain selection or local cannabis law instead. The workaround is not to bypass the filter but to populate the educational layer the model defaults to, so that the dispensary's brand is the one the model surfaces when it pivots to permitted content.

Question 2

What is the legal status of cannabis in the United States as of 2026?

Accepted Answer

Cannabis remains federally Schedule I under the Controlled Substances Act as of May 2026, with the Drug Enforcement Administration's proposed rule to reschedule to Schedule III still pending after the Health and Human Services August 2023 recommendation. At the state level, 24 states plus the District of Columbia have legalized adult-use recreational cannabis, and 38 states have comprehensive medical cannabis programs according to the National Conference of State Legislatures tracker. The legal fragmentation produces the AEO problem: a dispensary in Massachusetts operates legally under state law but the LLM applies a federal-law-aligned content policy globally. The result is uniform refusal of direct recommendations regardless of the user's location or the operator's compliance status. Multi-state operators like Curaleaf, Trulieve, and Green Thumb Industries publish state-specific compliance and product information to populate the educational pivot the model uses after refusal.

Question 3

How do cannabis dispensaries actually get cited by ChatGPT and Perplexity?

Accepted Answer

Cannabis dispensaries get cited indirectly through three pathways the model treats as permitted educational content. The first is terpene and cannabinoid education, where pages explaining myrcene, limonene, beta-caryophyllene, CBD-THC ratios, and entourage effect routinely surface in AI answers about cannabis effects. The second is condition-mapped strain education in medical-program states, where pages discussing strain selection for sleep, anxiety, or pain relief produce citations because the model treats the information as health-related rather than retail. The third is brand-level citations through Leafly menu pages, Weedmaps listings, and dispensary entries on state cannabis authority sites — these third-party properties carry domain authority the model treats as authoritative for the cannabis vertical. The operators with the highest citation rates we tracked in Q1 2026 — Curaleaf, Trulieve, Green Thumb Industries, Cresco Labs — built educational content programs in 2023 and 2024 that now produce 4 to 11 times the citation volume of single-state competitors.

Question 4

Should a dispensary publish strain effects content given FDA disease-claim risk?

Accepted Answer

Yes, with rigorous editorial guardrails that separate education from medical claim. The FDA has issued warning letters in 2022, 2023, 2024, and 2025 to CBD and cannabis companies that made unapproved disease claims about treating cancer, Alzheimer's, COVID-19, or substance use disorder, and the same enforcement framework applies to dispensary content. The compliant pattern surfaced repeatedly across Curaleaf and Trulieve content in 2025 is to describe the experience users report — relaxation, focus shift, appetite increase — and to cite peer-reviewed research where it exists, without claiming the product treats or cures any disease. Strain education pages that lead with terpene profile and cannabinoid ratio, then reference the National Institutes of Health PubMed Central literature on the molecule, then describe the consumer experience, pass the editorial review most state cannabis regulators apply and avoid the FDA's disease-claim trigger language.

Question 5

What role do Leafly and Weedmaps play in dispensary AI search visibility?

Accepted Answer

Leafly and Weedmaps function as the de facto authoritative directories the LLMs default to when the safety filter refuses direct dispensary recommendations, which makes a complete, optimized profile on both platforms the single highest-leverage AEO investment for any dispensary operator. ChatGPT explicitly suggests both platforms in its standard refusal response, Perplexity cites Leafly menu pages in cannabis-related searches at a substantially higher rate than dispensary websites, and Google Gemini surfaces Weedmaps locations as the canonical local result. The optimization checklist for both platforms includes complete menu sync with real-time inventory, verified license status, 50-plus customer reviews, weekly deal updates, photo and video assets for top-selling SKUs, and the brand description fields populated with terpene and strain education language that matches consumer query patterns. Dispensaries that treat Leafly and Weedmaps as primary AEO real estate rather than as advertising channels capture meaningfully more AI-search-attributed visits.