Using customized AI, marijuana retailers can analyze sales data in minutes

Just Released! Get realistic market forecasts, state-by-state insights and benchmarks with the new 2024 MJBiz Factbook member program, now with quarterly updates. Make informed decisions.

Image of person looking at charts on a laptop

(Photo by Kenstocker/

(This story is part of the cover package in the May-June issue of MJBizMagazine.)

For years, cannabis retailers have taken cues from more established industries, applying their best practices, systems and business processes.

However, in one increasingly important area, the industry is well ahead of other retail consumer segments: utilizing artificial intelligence to make business decisions.

And surprisingly, regulatory compliance is a big reason for that.

Analyzing track-and-trace data

State-mandated track-and-trace systems often draw widespread ire but also provide clear insights on so-called first-party data, or transactional information about when a product sold, the price and what else was in stock at the time of purchase.

Every cannabis point-of-sale system provides this data, and AI applications are helping turn these metrics into actionable insights.

“Cannabis is far ahead when it comes to first-party data capture for real interactions,” said Andrew Watson, a data scientist and CEO of Happy Cabbage Analytics.

“We should, as an industry, capitalize on that advantage.”

San Francisco-based Happy Cabbage certainly has. Its revenue-optimization platform is used in more than 400 marijuana stores.

Retailers pay $900 for the Happy Cabbage software suite, which can help operators train budtenders, predict product demand, target consumers, set prices and stock-keeping units (SKUs), prioritize discounts and perform other services.

“Happy Cabbage has always been about using data to create tools for retailers that help predict actions for them to take that will drive revenue and profitability and optimize inventory for maximized sell-through,” Watson said.

The company is among a new cadre of AI service providers for the cannabis industry, comprising a mix of startups and well-established players.

Cannabis inventory planning

Cannabis analytics company Headset records nearly $50 million in daily marijuana purchases.

The Seattle-based firm has utilized AI for years for internal operations, helping its analysts categorize, forecast and monitor various datasets to project category sales across 17 markets.

Its proprietary data, packaged through a few product offerings, can help retailers stock inventory, measure or plan promotions and, ultimately, drive more sales.

However, the output is only as good as the input, cautions Headset Chief Technology Officer Scott Vickers.

“You really need good data, which is what we have,” he said.

In early 2023, Headset started incorporating more large language models (LLMs) – AI systems that understand and can generate human language – to increase productivity, or the rate at which it could sort through enormous datasets and retail transactions.

Leveraging the power of LLMs and generative AI, Headset in September introduced Brand Beats, which allows customers to subscribe to brands and receive a weekly report highlighting key developments.

“What previously would’ve taken someone hours to assemble – digging through different dashboards and pulling out nuggets of insight – is now dropped right in their inbox every Monday morning,” Vickers said.

“We’re looking at doing very similar things for our retail customers.”

Cannabis recommendations

Pluggi, a cannabis-focused AI company with offices in San Francisco, London and Madrid, developed a chatbot that integrates directly with cannabis dispensary e-commerce menus such as those hosted by Dutchie and Jane Technologies.

More than 80 cannabis retailers across the country are using the service, which provides AI-driven product recommendations to shoppers within seconds.

“We do this by asking them three simple questions regarding the product type, flavor and effects that they desire,” Pluggi founder Wyatt Hahn explained.

The responses, garnered through online interactions, then are integrated into e-commerce platforms, cross-referencing data about terpene content, keyword product descriptions, strains and accompanying effects.

Pluggi bots are trained by consuming gobs of data, from frequently asked questions and ongoing product promotions to store hours and accepted payments.

“We’re all about bridging that gap between the in-store shopping experience – where you have access to a budtender who can assess your needs, make recommendations, address customer support issues – and the e-commerce experience, where traditionally that isn’t available,” Hahn said.

“We’re filling that space with AI.”

Pluggi charges retailers a $200 flat monthly subscription fee, which includes a customer-support chatbot.

The company’s target market is operators with two to five stores, according to Hahn.

He pegs the software’s return on investment at $4,600 per month, per store.

“We almost never lose a customer that’s in that two- to five-dispensary-chain range.”

‘Sales assistant for budtenders’

Another widely used cannabis retail-oriented app is Jointly, which since 2020 has helped track and guide the marijuana experiences of more than 500,000 consumers.

The goal, according to co-founder and CEO David Kooi, is to promote purposeful cannabis consumption, or identify reasons for use, such as relaxation, stress relief, sleep or creativity.

The Los Angeles-based company last year introduced Spark Pro, leveraging millions of data points on consumer behavior and buying patterns to help retailers curate specific products for shoppers.

“It’s like a sales assistant for budtenders,” Kooi said.

“It’s intended to bring some data to the conversation about what products somebody’s choosing.”

As of late April, about 25 retailers were using the upselling application, which costs $499 per month, per location.

“The approach we’re taking is that the AI is there to assist the budtender, make their job easier and make them more effective – not to drive the conversation.”

2024 MJBiz Factbook – now available!  

Exclusive industry data and analysis to help you make informed business decisions and avoid costly missteps. All the facts, none of the hype. 

Featured inside: 

  • Financial forecasts + capital investment trends 
  • 200+ pages and 49 charts highlighting key data figures and sales trends 
  • State-by-state guide to regulations, taxes & market opportunities
  • Monthly and quarterly updates, with new data & insights
  • And more!

Next-generation services

CannManage is among a new crop of startups offering cannabis companies AI services such as customized OpenAI GPT and targeted AI integration.

Founder Beth Myers and Chief Technical Writer Thomas Buckley have been utilizing generative AI to help automate and expedite the licensing process for clients, a cumbersome and expensive task in nearly every medical and adult-use marijuana market that often includes consultants, legal teams and other local business experts.

Since its August launch, Denver-based CannManage primarily has focused on automating certain aspects of technical writing for marijuana business licensing applications, incorporating plug-ins and other components to enhance capabilities.

But CannManage’s expansion plan is far more encompassing.

“We’re going to centralize them into one application software,” Myers said.

“I anticipate we’ll have a really good product to market here in the next two quarters.”

Myers, who has spent more than 16 years producing cannabis content, helping marijuana businesses secure licenses and managing retail operations, is excited about the next phase of AI integration in the industry.

“Every single one of us has a gem, and we’re running with it.”

Chris Casacchia can be reached at