Sarah Kabakoff is an award-winning forward thinker who deployed emerging technologies working at Toast until going to work in cannabis. As Director of Enterprise Solutions Engineering at Toast, she identified gaps in restaurant technology. With her background in Point of Sales (POS) and her eye for innovation, it made sense to join an e-commerce company like Dutchie.
She joined them in late 2020 before they made their POS acquisitions of LeafLogix and Greenbits. There, she hired from within to build camaraderie and gave everyone a clear path for growth and leadership in an equitable organization. Eventually, she was promoted to VP of Sales and during that first year they captured 10% of the cannabis market in sales, closing hundreds of locations a month. Then she met the team at Genetica and decided to take a leap.
Getting in With Genetica
At Genetica, Kabakoff was brought in as a fractional Chief Revenue Officer, helping build systems and give advice before coming on full-time when her position changed to President and she oversaw product engineering and revenues strategy. She wanted to work on innovative solutions and be able to oversee the product team.
“I need to be the dumbest person in the room for all my people that are working with me, and then we’ll be successful,” Kabakoff insisted.
She liked the idea of a system that could run everything and give alerts on what’s happening in a cannabis business in real-time when that information was needed.
“Historically, businesses have had to go do a bunch of work and have analysts, to pay [them] to build new reports and everything else,” said Kabakoff.
She really dug the AI, which still innovates in the hospitality sector, and is rapidly growing as a global cannabis business. They already power systems in Canada, Europe, and Australia. They have big plans in the works for the United States, too.
AI in Action
Using a Language Learning Model (LLM) is just one of the tools, but they use all kinds.
“LLM is just one flavor of AI. And we do utilize those to help communicate to humans. But LLM’s aren’t good at math,” Kabakoff advised.
They are moving into voice very soon, and currently have AIs that do different things like text-to-sequel and text-to-cipher (a Python language that creates queries and then sequel for more basic things).
“We put a set of tools in front of the AI, and the AI picks up a tool based on the user’s input, and then it takes that tool and goes to get the answer,” she said.
They also build vertically-specific agents. Being able to identify purchasing patterns with the AI, then look at the sales velocity of specific items allowing businesses to create the right bundling or special in order to move an item properly. It ensures they won’t see a drop in revenue which is key to their success.
“You can’t just tell the machine cannabis,” Kabakoff explained. “You have to be able to understand what all the relational terms to ‘cannabis’ are.”
They utilize ontology which helps the machine understand what the data points are and to be able to navigate that.
“It’s like loading the knowledge in your brain to the AI to allow it to understand what it’s looking at and how to navigate,” she said.
Genetica is very good at making the brain of the AI, and being able to load that and use things like Model Context Protocol (MCP) servers to teach it how to do something.
“Here’s what it is. Here’s how you do it. Combine the two, and then use the tools in front of you to take actions,” Kabakoff instructed.
They look at everything surrounding a purchase: the user, the product, things they’re bundling, their history. All the factors. Even the complexity of a type of product to understand what the right mix is that sells better than others or has more repeat customers.
Disruption Versus Displacement
AI will displace some jobs, but mainly for people who don’t engage in it.
“AI is going to be horizontal, like electricity. It’s going to be in everything,” Kabakoff warned.
She believes that the jobs that will be affected are white collar jobs: executive positions, mid-level managers, software engineers and analysis roles. High level positions and executors will retain strategy while mid-level positions will start to be eliminated. “AI does it all. If I have the right framework, I let the agents run overnight and build, come back in the morning and critique it. They do the work.”
“People want to know about their business when they wake up in the morning and they’re on the toilet,” quipped Kabakoff.
Instead of needing analysts, or waiting on someone to provide a specific report, they can ask in natural language for that dashboard to be built and it will be.
Genetica provides the ability to be able to understand the perception about your business in real time via agents monitoring everything—from Yelp, Google and Reddit threads to the deepest recesses of the internet. It results in less billable hours, less components, and real-time analytics which makes for a more seamless, profitable endeavor.
“No more putting in a ticket or trying to find someone to figure it out or build pivot tables,” Kabakoff said. “Now you can just have this do it for you.”
