Artificial intelligence panel

According to the two executives, Montreal enjoys a unique position in the field of AI. Not only does the city have a large pool of researchers, but it also offers a more human-centered approach than the one often associated with Silicon Valley.

Valérie Pisano even compared local AI leaders to the “Céline Dions of artificial intelligence,” citing Yoshua Bengio, Hugo Larochelle, and Joëlle Pineau in particular.

“We don’t pretend to have talent—we actually have it,” she said. “And we do it differently. Here, we don’t subscribe to the ‘move fast and break things’ philosophy.”

This more “sociotechnological” approach, as she put it, is also attracting foreign delegations seeking an alternative to dominant American models.

Joëlle Pineau also believes that Cohere’s decision to develop AI primarily for businesses and governments allows the company to remain grounded in practical applications.

“We’re in an era where there’s a lot of hyperbole surrounding AI,” she explained. “In most cases, it doesn’t yet fully meet people’s needs. So I find it interesting to be immersed in real-world problems rather than thinking we’ve invented some magic solution that will fix everything.”

Long recognized for its academic research, Canada has often been criticized for its difficulty in turning discoveries into strong, successful companies. The two speakers, however, believe that a cultural shift is now underway.

Valérie Pisano notably emphasized the emergence of the “researcher-entrepreneur,” a profile that Mila is now seeking to support more actively through various programs, including the Venture Scientist Fund, created through a partnership between Mila and Inovia Capital.

“Researchers now want to take their work and bring it into the real world,” she explained. “They want to build their own companies around their values and priorities.”

The challenge also lies in retaining talent in an environment where American companies have considerable financial resources. “We cannot take this for granted,” Valérie Pisano warned. “We need to invest, build bridges and create projects. We have to go all in.”

For her part, Joëlle Pineau believes that Montreal still has several competitive advantages despite salaries that are sometimes lower than in the United States. “People come for the right reasons,” she said. “They want to be part of something meaningful and live in a city where they actually want to stay.”

She acknowledged, however, that immigration policies can still create obstacles when it comes to recruiting certain international talent.

Cohere’s rapid expansion also illustrates Canada’s ambition to move from the laboratory to the global market. Founded in 2019, the company now has offices in Toronto, Montreal, San Francisco, New York, London and Paris. It has also recently announced strategic acquisitions in Germany.

According to Joëlle Pineau, this growth is driven in particular by three factors: data security, technological sovereignty and cost control. “Many companies and governments are looking for solutions that allow them to keep their data locally and reduce their dependence on American technologies,” she explained.

The costs associated with the large-scale use of major American models are also becoming a significant issue for many organizations. “Token costs [the units of data processed by AI models] can quickly get out of hand,” she noted, adding that Cohere is working to develop models that are more efficient and less expensive to deploy.

Beyond technological considerations, the two speakers issued a warning to leaders: inaction now risks costing more than a bad decision.

Joëlle Pineau drew a parallel with the consultations Ottawa is conducting on its future national strategy for artificial intelligence. According to her, many organizations fall into the trap of wanting to wait for the perfect plan before taking action. “There won’t be a perfect strategy,” she stated. “Just having one and moving forward is what matters most.”

Valérie Pisano, for her part, believes that companies will need to thoroughly rethink how they manage change. Traditional decision-making cycles will no longer suffice in an environment where technologies evolve from one month to the next.

“We normally prefer to do very thorough work, reach a conclusion, and then execute the plan,” she explained. “But AI won’t allow us to do that. We’ll have to learn to test, adjust, start over, and accept a certain amount of discomfort.”

The message is clear: while organizations are still hesitating, global competition is moving forward rapidly.