The Power List: Ivan Zhang, Aidan Gomez & Nick Frosst are creating a smarter, friendlier chatbot

AI leader Cohere springs from the brains of three young men who are fascinated by the questions natural language processing poses

H.G. Watson
Content image

AI Trailblazers

No.1: Ivan Zhang, Aidan Gomez & Nick Frosst

READ: The Power List: AI top 10

If you’ve communicated with your bank lately—or a retail store or internet provider—the person on the other end of the chat probably wasn’t, well, a person. Most of our first interactions with AI won’t be via sentient spaceships or futuristic droids, but rather the humble chatbot. And one of the leaders in this space is Cohere, a Toronto startup founded in 2019. Cohere produces chatbots by using natural language processing, a field of artificial intelligence focused on how computers communicate with us. They train large neural networks—computer systems modelled after the human brain—on terabytes of data to teach them how to communicate in a human way. And when those networks have absorbed all that data, companies can use them for far more than chatbots. Need to sort huge amounts of data? Come up with a business plan? Write social media copy? Cohere’s technology can do all that and more. 

MORE: See who made the 2023 Maclean’s Power List

The company springs from the brains of three young men who are fascinated by the questions natural language processing poses. In 2017, Aidan Gomez, a University of Toronto alumnus, co-authored an important paper on how transformers— neural networks that learn how to understand information with context—could revolutionize how artificial intelligence understands language. Ivan Zhang, meanwhile, was a U of T dropout who often hung out at Google Brain, where Gomez worked. Eventually, the two of them decided transformer technology was the key to natural language processing, and they built a company around it. Nick Frosst, meanwhile, got started in the field after a casual chat with a computer science student during his part-time job at Snakes and Lattes led to a research position. He joined the Cohere team in 2020. 

Chatbots went mainstream in November of 2022, when OpenAI—a Microsoft-backed artificial intelligence company and Cohere’s biggest competitor— got a lot of press when it launched ChatGPT. The chatbot made headlines for its ability to mimic human conversation: one ChatGPT-powered search engine even told a reporter that it wanted to be human. Cohere’s technology works similarly, but its founders decided to focus on business applications over public displays— Gomez recently told Reuters that they plan on launching a dialogue model like ChatGPT that users could use to generate content—though within stricter parameters. The model will still be used for business, and not for the more creative uses ChatGPT users have attempted. 

RELATED: Why this professor supports his students using ChatGPT to write papers and exams

While OpenAI gets the buzz, Cohere is getting action behind the scenes. It’s focusing on advanced business applications over splashy public gimmicks. By 2022, the company had grown from the three founders and three engineers to nearly 100 people working in Canada, the United States and the United Kingdom. Last year, it raised $125 million in Series B funding, and earlier in 2023, it scored a major coup when Martin Kon, formerly YouTube’s chief financial officer, joined the company as president and chief operating officer. Gomez has said that they expect this to be a breakout year for landing new enterprise clients. Now Cohere is in talks for a new funding round that might increase the company’s valuation to $6 billion. A vote of confidence came in the form of a major investment by Salesforce Ventures, who chose Cohere as one of four companies to get part of $250 million funding for generative AI technology.

Check out the full 2023 Power List here

This article appears in print in the March 2023 issue of Maclean’s magazine. Buy the issue for $9.99 or better yet, subscribe to the monthly print magazine for just $39.99.