The Fork in the Road
When ChatGPT 3.0 launched, I dove into prompt engineering like it was the only game in town. But models evolve fast, and these days even a clumsy prompt can get polished on the fly. Prompting is still important, but I wanted something bigger.
So here I am, shifting my blog from prompts to my latest obsession: building AI automations that solve messy, everyday problems.
Where I’m Coming From
By day, I’m a full-charge bookkeeper in the cannabis industry. Not exactly the most glamorous title, but hear me out. If you picture cannabis as a shiny new industry full of slick tech, you’d be wrong. It’s mostly paper, clunky processes, and software vendors who refuse to work with us. (Yes, Netsuite, I’m calling you out.)
Here’s the funny part: I don’t even smoke. I don’t know brands, strains, or what’s “fire.” My lane is the money—and the systems that manage it.
Why This Industry?
Cannabis businesses spent years in the shadows. Federal law still lists it as a Schedule I drug, and legality depends on which state you’re standing in. That makes operations inefficient and often ignored by mainstream software companies.
To me, that screams opportunity. I don’t just want to bring digital band-aids to the table—I want to serve AI-powered solutions that actually change how things run.
My Starter Toolkit
I’ve narrowed my stack to three main tools:
- ChatGPT – my sounding board, idea refiner, and automation assistant.
- Google Sheets – the workhorse for shaping data.
- Zapier – the connector that makes everything talk to each other.
The First Build
Here’s the first automation I’m rolling out:
- Pull data from our POS system.
- Clean and shape it in Google Sheets.
- Generate a journal entry.
- Send it through Zapier to QuickBooks Online.
The result? The books update automatically for each dispensary. Right now I manage three locations, and more are opening by year’s end. The POS already handles inventory, so I don’t waste time entering every vendor payment into QBO. Instead, I let the process run monthly in the background.
The Next Challenge
Administrative expenses are still a headache. But even there, QBO has features I’m finally putting to work—like forwarding invoices straight into the Accounts Payable queue (they call it “Receipts”) and using feeds to cut down manual entry.
Where This Is Going
This is just the start. My goal is to document the wins, the mistakes, and the lessons as I go—so if you’re curious about how AI can handle real-world business workflows (especially for an industry people love to ignore), you’ll see it unfold here in real time.