There’s a Robot for That: PotBot and BrainBot Simplify Medical Marijuana
PotBotics is one of the most advanced medical technology companies in the cannabis industry, but the problem it aims to solve is simple: removing the guesswork from medical marijuana. Right now, patients go through what co-founder David Goldstein calls a “silly process” to find cannabis treatments. A doctor will typically listen to a patient’s symptoms, recommend medical marijuana as a treatment, and hand them a prescription for a medical marijuana card. Typically, the conversation between the doctor and patient ends there. It’s now up to the patient to go to a dispensary and ask budtenders, who often have no medical background or training, which strains or products would work best for their illness. Some patients may dig for hours online, searching for anecdotes and limited research about which types of marijuana would help with their particular ailment. PotBotics wants to change this conversation around medical marijuana by developing technologies that will help doctors and budtenders alike in recommending strains that are best suited to the patients’ needs.
The company, founded by Goldstein and his father, Baruch Goldstein, has three technologies currently in development: PotBot, a virtual budtender, BrainBot, a wireless electroencephalography (EEG) helmet for measuring patient response to marijuana, and Nanopot, a genetics evaluation system for cannabis strains. The company, which has offices in both New York City and Palo Alto, launched in 2014 and is pushing PotBot and Brainbot to market this year.
While recommending a strain of marijuana based on flavor and smell may be appropriate in the recreational market, both PotBot and BrainBot are based on the key compounds that make marijuana a medical treatment: cannabinoids. “Cannabinoids are the medicinal backbone of cannabis,” Goldstein said. “They are the main ingredient – the main compound, really – that attributes to therapeutic relief, whether that be CBC for anti-anxiety properties or CBD for anti-inflammatory properties or THC for helping improve appetite. All of these cannabinoids are really the points that we wanted to engage with patients.” Both PotBot and BrainBot can currently help identify the values of six cannabinoids, and PotBot links those values to strain names. While there are over 1,000 strains on the market, the technologies are focused on a select number of “authentic strains”, or strains that are parents to the many varieties made available through genetic engineering.
Goldstein imagines that in the near future, medical marijuana patients will visit with their doctor, who will use BrainBot to test the effects of a few cannabis strains on their brain and provide a recommendation for cannabinoids based on how they react to the different levels in each one. Then, the patient will go to a dispensary and pull up their patient profile on PotBot, which will give them information on the recommended strains that match their ideal cannabinoid profile. The store’s budtender will assist the patient in finding products derived from that strain.
The two technologies can also work independently, though – a BrainBot patient can walk into a store without a PotBot and ask a budtender for a strain with the cannabinoids recommended by their doctor, or a medical marijuana patient could use PotBot alone to search for a strain that’s best for them.
While BrainBot is based on EEG technology, which has been around for nearly a century, PotBot is built using Semantic Web, which has only been in development for a handful of years, with companies like Google and IBM Watson leading the charge. As Goldstein explained, traditional artificial intelligence linked data through “if, then” statements. So, using a system built on artificial intelligence, it would know that if someone typed in “insomnia”, it would search the system’s manually entered data for “CBD”, a type of cannabinoid. With Semantic Web, data is linked through triplets, which, in laymen’s terms, are like triangles with each point linked to a type of information. Using this system, PotBot can find information in many studies found online, rather than just those that are manually entered into the system. “You are able to pull correlations from the word insomnia that go much deeper than just an input-output architecture,” Goldstein explained.
BrainBot is wrapping up its pilot program with 400 patients this month. Goldstein said that one of the patient’s words alone proved the company’s concept. “She told me, ‘What took me four months of research, I just learned in five minutes.’ That just validated everything we were trying to achieve,” he said.