How a cannabis DTC brand won #1 for

A category AI is built to avoid

Cannabis is one of the hardest categories in AI search. Models answer cautiously and state by state, citations are scarce, and most brands never appear in the response at all. 1906, a DTC cannabis brand, needed net-new customers from AI and organic search, but the usual playbook doesn't scale in a space this restricted. The real question was whether an LLM could be made to cite a cannabis brand at all.

Cannabis is the category AI is built to be careful about, which is exactly why almost no one wins it.

Citations where competitors have none

We ran 1906 through the T.R.U.S.T. framework and focused entirely on non-branded, transactional and commercial demand ("cannabis pills," "THC packets," "edibles pills," "arousal edibles"), so every gain was a net-new buyer. We built article-first content engineered to be citable within the constraints models apply to cannabis, seeded semantic mentions on Reddit and Quora to feed the sources LLMs pull from, and did the technical work to make the pages legible to Google AI Overviews, ChatGPT, Perplexity, Gemini and Copilot.

We built the citations in a category where competitors have none.

#1 for the query that matters most

In three months, 1906's AI citations grew 95%, led by 59 citations in Google AI Overviews with gains across ChatGPT, Gemini and Copilot. The brand took the #1 position for "top THC pills," a high-intent, high-value query, and lifted its commercial and transactional top-10 keyword footprint 19% in the first 60 days, ranking for terms like "cannabis pills" (position 25 → 3, +53 monthly visits) and "THC packets." Every one of those keywords is net-new, non-branded demand.

#1 for "top THC pills," in the category AI works hardest to stay neutral in.