Read five dispensary blogs in a row and you start to notice something. "Whether you're a seasoned connoisseur or a curious newcomer." "A journey through the world of terpenes." Different shops, different states, the same sentences. A lot of cannabis content is now written by chatbots, and it shows.
It's easy to see why it happens. Cannabis businesses can't buy ads on Google or Meta, so search traffic and content do most of the marketing work. A blog post a day sounds like a plan, and a chatbot can produce one in thirty seconds. But readers (and search engines) have caught on.
Why it hurts more than it helps
It all sounds the same. The whole point of a dispensary blog is to give someone a reason to choose your shop over the one three blocks away. Generic AI copy does the opposite: it erases the budtender who actually knows the menu, the grower you work with, the regulars who keep coming back.
Search engines are pushing back. Google's spam policies now target "scaled content abuse", meaning large volumes of pages produced mainly to rank rather than to help anyone, and that applies however the pages were made. A hundred near-identical strain guides is the textbook example. Pages like that tend to sink, and they can drag the rest of a site down with them.
It invents things. Chatbots fill gaps with confident guesses, and in cannabis that's a real problem. A post that claims a strain "relieves anxiety" or "helps with insomnia" is a health claim, and most state rules restrict those claims in cannabis marketing. The model doesn't know your state's rules. It doesn't know your menu either, and it'll cheerfully describe a strain you've never stocked!
How to spot it in your own content
Start with the phrases. Language models lean hard on a small set of words and stock expressions; "a symphony of flavors" is a favorite. Any one of them is fine in isolation. A post full of them reads like a template. There's a public list of the words AI models overuse, updated as the models change, and it's a quick way to audit a draft.
Then look at the shape. AI copy has a rhythm: an opener that restates the title, three tidy sections of equal length, bullet points that all start the same way, and a neat summing-up at the end. Real writing is lumpier.
Finally, look for specifics. Does the post mention anything only your shop would know? A price, a grower, a staff pick, a question customers actually ask at the counter? If you could swap your logo for a competitor's and nothing would change, the post isn't doing its job.
How to fix it without hiring an agency
You don't have to give up AI tools completely. Use them for outlines or for tidying grammar. Then put the shop back in.
- Interview your budtenders. Ten minutes of "what do people ask you about this strain?" gives you more usable content than any prompt. Quote them by first name.
- Write about your menu, not the category. "Our three favorite indicas this month and why" beats "The Ultimate Guide to Indica Strains" every time.
- Use your own photos. Product shots from your counter, not stock images of a generic leaf.
- Cut the throat-clearing. Delete the first paragraph of any draft and see if the post still works. It usually does.
- Read it out loud. If it doesn't sound like something anyone on your staff would say, rewrite it until it does.
- Publish less. One real post a week will do more for your shop than seven generic ones.
Sound like your shop
Customers come to a dispensary for a person who knows the product. Your website should sound like that person. Chatbot copy is cheap to produce. It also costs you the one thing a small cannabis business has that the chains don't: a voice people recognize and trust.
You might also like
Leave A Reply