Cannabis is often positioned as an alternative to mainstream medicine. To keep up, some companies are now trying to fine-tune their offerings to remove the unpredictability long associated with marijuana.

From an electronic pot-recommendation program launched last month by Columbia Care to DNA analysis by startups such as Endocanna Health, the industry aims to create products akin to personalized or precision medicine, optimized based on genetics, lifestyle or even what kind of experience users want to have. Algorithms that can do so are seen as a holy grail for the cannabis industry, quelling the fears of first-timers and satisfying connoisseurs alike.

“This kind of technology could be a potential game-changer for plant-based medicines, including the cannabis industry,” Samoon Ahmad, a professor of psychiatry at NYU Grossman School of Medicine and co-author of “Medical Marijuana: A Clinical Handbook,” told me in an email exchange. “Going back more than 100 years, researchers have been frustrated by how unpredictable a user’s experience with different samples of cannabis can be.”

Columbia Care, a New York-based multistate operator with expertise in medical cannabis sales, saw a big gap in the market when it came to matching customers to different plant strains. Jesse Channon, the company’s chief growth officer, said it took about a year to develop its recommendation program, known as Forage.io.

Forage, which resembles an app and is available online or on kiosks in its stores, asks users questions as simple as “How do you want to feel today?” A digital “mood ring” helps the customer indicate what they want to experience, and then the questions determine what kind of format they want to consume it in.

“It’s like a guided discovery experience,” Channon said. It can take sales staff years to master the kinds of recommendations that Forage can now help all employees make quickly in its dispensaries. The program, which the company plans to roll out to almost 100 dispensaries, has also served as a conversation-starter with customers, he said.

Endocanna Health, based in Burbank, California, has gone even deeper into personalized cannabis, with tests that analyze people’s DNA. CEO Len May said in a phone interview that the company currently tests for around 162 factors that could show genetic predispositions to different interactions with cannabis.

That can help users avoid the kind of negative outcomes that might turn first-timers off cannabis for good, he said. For example, around half of the U.S. population has a gene called FAAH that will influence whether too much THC, the active ingredient in marijuana, might trigger an anxiety event.

“If you have a genetic predisposition to anxiety, it doesn’t mean you shouldn’t consume,” May said. “It just means you consume differently.”

RealSleep, based in Los Angeles, is a company that has used wearable devices like Fitbits or Apple watches to help people sleep better with cannabis. Jointly has an app designed to help people track their own cannabis experience based on 15 factors like dose, hydration and hunger levels so they can customize or improve it the next time around.

The market is in its infancy, but given how unpredictable the interactions between individuals and cannabinoids are, the programs with the deepest understanding of personalization may emerge as the leaders.

“Efforts to match people to strains without understanding their genetics are incomplete,” cautioned May. For some users, conventional wisdom — like Sativa strains for uplift and Indica strains for more toned-down experiences — may not apply.

“We have millions of genes and there are over 400 different components in cannabis plants,” he said. “We want people to have a personalized experience, not just guess.”