AI agents tools promise to get out of some intoxication from the daily workflows, but most organizations are still hesitant to adopt them, a pressure is a matter of concern: data protection. Business secrets, large enterprises with companies and government agencies in highly regulated industries have thought about bringing AI equipment out of the concern that their – or worse, their customers – data – data – data can be unknowingly compromised, or used to train the foundation model.
Canadian AI firm Kohere with its new AI agent platform aimed at reducing concerns that have been dubbed to 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 as good as they have access,” said Kohere’s co-founder and CEO Nick Frost, during a demo in the answer. “If we want llms to be as useful as possible, they will have to reach that useful data, and means they need to be deployed in (customer’s) environment.”
Instead of using enterprise cloud platforms such as Azure or AWS, COHERE says that it can install the answer on the private infrastructure of an organization so that it never looks or interacts with the customer’s data. North can run on an organization’s on-primesies infrastructure, hybrid clouds, VPC or air-gapped environment, said the Frost.
He said, “We can literally deploy on a GPU in a closet that may be with them somewhere,” he explained, saying that the answer was designed to run on something as two GPUs.
COHERE claims that the north also includes safety protocols such as granular access control, agent autonomy policies, continuous red-teaming and third-party security testing. And, it meets international compliance standards like GDPR, SOC-2 and ISO 27001.
More than private deployment
Cohere, which has raised $ 970 million so far, recently evaluated $ 5.5 billion, said that it has already piloted to north with some customers such as RBC, Dell, LG, LG, ENSIMBAL Health Partners, and, as TECRunch had reported Palantir last year.
Techcrunch event
San francisco
,
27-29 October, 2025
North showed several AI agents platforms just out of the box. Its main features are chat and search, which answers users for customer support inquiry; Complain the meeting tape, write marketing copy, and access information from both internal resources and web. Frost stated that all reactions include quotes and “logic” chains of thought so that employees can audit and verify the output.
Chat and search functions are powered by existing cotton techniques, such as commands (its family AI model’s family), and compass (its multimodal search tech stack). Frost stated that North is powered by a version of its command model that is trained for enterprise logic.
“This is just beyond Q&A and involves working for you. Therefore, (North) has a group of asset construction. It can make a table, it can make a document, it can create slideshows. It can make a group of market research,” Frast said.
It is worth noting that in May, the Vancouver-based platform acquired Otogrid, which develops enterprise equipment to automate high-level market research.
Like other AI agents platforms, North can connect to existing workplace tools such as Gmail, Slack, Cellsforce, Outlook and Layer, and can be unified with any model reference protocol (MCP) server to access industry-specific or in-house application.
“As you create confidence by chatting on the model, there is a smooth transition that is between using it as an increase to use it as an increase,” the Frost said.
