Cohere’s new AI agent platform, North, promises to keep enterprise data secure

AI agent tools promise to siphon out some of the drudgery from daily workflows, but most organizations are hesitant to adopt them yet, harboring a pressing concern: data security. Large enterprises with trade secrets, companies in highly regulated industries, and government agencies have thought more than twice about bringing in AI tools out of concern that their — or worse, their customers’ — data could inadvertently be compromised, or used to train foundation models. 

Canadian AI firm Cohere is taking aim at alleviating those concerns with its new AI agent platform dubbed North, which promises to enable private deployment so that enterprises and governments can keep their and customers’ data safe behind their own firewalls. 

“LLMs are only as good as the data they have access to,” Nick Frosst, co-founder and CEO of Cohere, said during a demo of North. “If we want LLMs to be as useful as possible, they have to access that useful data, and that means they need to be deployed in [the customer’s] environment.”

Instead of using enterprise cloud platforms like Azure or AWS, Cohere says it can install North on an organization’s private infrastructure so that it never sees or interacts with a customer’s data. North can run on an organization’s on-premise infrastructure, hybrid clouds, VPCs, or air-gapped environments, Frosst said. 

“We can deploy literally on a GPU in a closet that they might have somewhere,” he explained, adding that North was designed to run on as few as two GPUs. 

Cohere claims North also includes security protocols like granular access control, agent autonomy policies, continuous red-teaming, and third-party security tests. And, it meets international compliance standards like GDPR, SOC-2, and ISO 27001.

More than private deployments

Cohere, which has so far raised $970 million, most recently at a $5.5 billion valuation, said it has already piloted North with some customers such as RBC, Dell, LG, Ensemble Health Partners, and, as TechCrunch reported last year, Palantir. 

North mirrors many AI agent platforms right out of the box. Its chief features are chat and search, which let users get answers to customer support inquiries; summarize meeting transcripts, write marketing copy, and access information from both internal resources and the web. Frosst added that all responses include citations and “reasoning” chains of thought so employees can audit and verify the output.

The chat and search functions are powered by existing Cohere technology, like Command (its family of generative AI models), and Compass (its multimodal search tech stack). Frosst said North is powered by a variant of its Command model that is trained for enterprise reasoning.

“It goes beyond just Q&A and gets into doing work for you. So, [North] has a bunch of asset creation. It can make tables, it can make documents, it can make slideshows. It can do a bunch of market research,” Frosst said.

It’s worth noting that in May, Cohere acquired Ottogrid, a Vancouver-based platform that develops enterprise tools for automating high-level market research. 

Like other AI agent platforms, North can connect to existing workplace tools like Gmail, Slack, Salesforce, Outlook and Linear, and integrate with any Model Context Protocol (MCP) servers to access industry-specific or in-house applications. 

“As you build confidence by chatting to the model, there’s like a smooth transition that happens between using this as an augmentation to using it as an automation,” Frosst said.