Skip to main content
Open this photo in gallery:

Left to right: Cohere co-founders Ivan Zhang, Nick Frosst and Aidan Gomez.Fred Lum/the Globe and Mail

A Toronto artificial intelligence startup, backed by some of the leading experts in the field globally, has raised US$40-million as it prepares to formally launch its technology to help make it easier for humans to talk to machines.

Cohere Inc., co-founded by protégés of AI luminaries Geoffrey Hinton and Jeff Dean, said the funding was led by Index Ventures, a venture firm with headquarters in London and San Francisco that has also backed Slack Technologies Inc., Sonos Inc., Robinhood Markets Inc., and Etsy Inc. Other investors include venture capital firms Section 32 and Toronto’s Radical Ventures - which wrote the original cheque to fund Cohere - and renowned AI experts including Mr. Hinton, Fei-Fei Li, Pieter Abbeel and Raquel Urtasun.

The startup says its natural language processing (NLP) software provides a richer understanding of human language, including semantics, sentiments and tone. As such, it improves on existing software that powers machine-human interactions such as online conversations between consumers and customer service chatbots. The idea is to provide the same quality of NLP models to a wide array of customers that only a handful of internet giants currently use.

Cohere, founded in 2019, came out of “stealth” mode in May by revealing that it has been working with a handful of test customers, including Toronto chatbot service provider Ada Support, plugging them into its software over the internet.

Cohere builds on a pioneering 2017 paper co-written by chief executive officer Aidan Gomez (a one-time intern in Mr. Hinton’s Toronto lab), when he worked for Google. The paper invented a concept known as “transformers,” which improved the efficacy of NLP technology that use AI algorithms to generate language. Mr. Gomez and Cohere co-founder Nick Frosst, another ex-Google researcher who previously worked with Mr. Hinton, wanted to translate gains in NLP from the research sphere to practical applications and set out on their own to do that. The third co-founder is fellow Toronto-based AI researcher Ivan Zhang.

Mr. Gomez said in an interview his 40-person company is planning for its product to be generally available this fall and has been adding sales and marketing staff for the formal launch. “We feel ready now to open things up more broadly,” Mr. Gomez said. “We really want to make it useful to everyone.”

Index Ventures co-founder Mike Volpi said in an interview he was impressed Cohere’s leaders “had a real focus on building a usable product that developers could just take and run with” for specific uses, such as helping writers draft articles, powering chatbots and helping with content moderation on websites.

Unlike some AI companies that generate a lot of buzz and raise big funding but are still more experimental and research-oriented, Cohere “was a little further away from academia and more focused on productizing this technology which I really liked a lot,” Mr. Volpi said. “Few companies have the discipline to say, ‘We are going after precisely these things which are practical and monetizable today.’”

Mr. Volpi added the team at Cohere was one of the few in the world, alongside AI specialists at Google and Elon Musk-funded non-profit OpenAI, with the skills to develop the kind of next-generation NLP technology Cohere is selling.

Cohere is the latest in a string of Toronto AI startups founded by University of Toronto professors-turned entrepreneurs to raise sizable financings this year. Those include Ms. Urtasun’s autonomous driving developer Waabi Innovation Inc. and Brendan Frey’s Deep Genomics, which uses AI to discover new drugs, and semiconductor developer Untether AI Corp., co-founded by Martin Snelgrove.

Your time is valuable. Have the Top Business Headlines newsletter conveniently delivered to your inbox in the morning or evening. Sign up today.

Your Globe

Build your personal news feed

Follow the author of this article:

Follow topics related to this article:

Check Following for new articles

Interact with The Globe