The Canada Nation

Your Trusted news Source

Cohere, Open Text to pitch Canadian-made AI agent tools to enterprises

Cohere, Open Text to pitch Canadian-made AI agent tools to enterprises



Open this photo in gallery:

Cohere co-founder and CEO Aidan Gomez speaks at the Milken Institute’s Global Dialogues Toronto summit, on Sept. 14.Sammy Kogan/The Globe and Mail

Cohere Inc. and Open Text Corp. are partnering to sell artificial intelligence tools to businesses and governments, as some companies are struggling to see a return from AI, while looking for alternatives to U.S. tech.

The two Canadian companies are working together to power agents, which can access data and use other applications to carry out work. Technology from Open Text will surface relevant internal data to serve as context for Cohere’s AI models and North platform to complete tasks.

“Agents are only as effective as the data, context, and governance behind them,” said Shannon Bell, chief digital officer at Open Text.

The companies said the joint offering, which will be available in early 2027, will appeal to governments and firms in regulated industries that require more control over their data and have higher security needs.

Joelle Pineau, Cohere’s chief AI officer, said: “There is a real appetite for solutions that are built outside of the U.S. that have guarantees of sovereignty.”

AI firms’ calls for co-ordinated slowdown amounts to ‘cartel’ behaviour, Cohere CEO says

Since its founding in 2019, Toronto-based Cohere has targeted corporate and public sector clients as opposed to consumers. Its large language models can run on clients’ own infrastructure, allowing them to have more control, the company has said. As conversations about digital sovereignty have flared up in the past year and governments and businesses look to decrease reliance on U.S. tech, Cohere has leaned even harder into pitching itself as an alternative.

Waterloo, Ont.-based Open Text, meanwhile, has a suite of tools for governing and managing data, and counts some 120,000 enterprise clients globally, including in health, finance and insurance. Customers are “in a good position to unlock that data,” Ms. Bell said.

But wringing value from AI tools is still a challenge for some companies, and studies have shown the results are mixed so far. A survey published in August by global consultancy McKinsey & Company polling 1,719 people in 97 countries found that 80 per cent of respondents said AI was improving their individual productivity, and half said it helped them make better decisions. But only 37 per cent said AI was making at least some contribution to the bottom line, which was unchanged from last year despite more companies adopting the technology.

Opinion: AI tech bros are self-serving and slimy. But we must heed them anyway and pause development

A March, 2026 working paper from the National Bureau of Economic Research in the U.S. found that chief financial officers perceived the productivity gains from AI to be substantially higher than the actual effect. That’s not unusual for transformative technologies, however, which are “widely viewed as important well before their effects are fully reflected in measured productivity,” according to the study.

Some companies still have a 2023 mindset of AI, too, and treat the technology as a simple chatbot instead of exploring the full possibilities of agents, Ms. Pineau said. Doing so requires governance and monitoring strategies to ensure that agents are working properly, which takes time to implement.

The McKinsey study noted AI tool expenses are pushing about 20 per cent of respondents to limit their usage, and cost constraints were reported for standard chatbots, agents and coding tools. Still, the report noted 60 per cent of respondents expected their companies to boost AI investments over the next year.

Ms. Pineau said that companies have realized that they don’t need to use the most powerful and expensive model for every task. “We’re seeing a large part of the workflows shifting from some of the large models built by OpenAI, Google, Anthropic and others, towards the types of models we are building at Cohere that tend to be more tuned to enterprise workflows, but also much more efficient.”

The partnership between Cohere and Open Text is not exclusive, and both will continue working with other partners. No financial details of the partnership were disclosed.