
What happened
Cohere Inc. unveiled North 2, an overhaul of its enterprise AI agent platform, with a redesigned orchestration system, token spending caps in North Admin, and Slack, GitHub, SharePoint and OneDrive connectors.
Why it matters
Cohere says enterprises had accepted giving up security, capability or cost control when deploying agents, so per-user and per-agent token caps and agent-level guardrails are aimed at that tradeoff.
What to watch
North 2 is model-agnostic, so value hinges on whether customers bring outside models and on Cohere's Nvidia work; it lands under three weeks after Cohere's merger agreement with Aleph Alpha GmbH.
WHO IT HITSEnterprise IT and platform teams running AI agents internally are the main audience, since North 2 tracks token use down to individual users and agents and lets administrators assign models, including ones brought from outside Cohere. Security operations teams are another, as Bell Cyber's use of North for agentic AI shows.
Summaries like this, in your inbox every morning.
North 2 is an upgrade rather than a debut: North has been generally available since August 2025, and Cohere says a year of large production deployments across industries from finance to the public sector shaped what went into it. That experience shows up in the design choices — agents built from a simple prompt can be shared across a company so other teams do not rebuild the same workflow, each agent keeps memory between sessions, and documents every agent should reach sit in shared libraries.
The orchestration system runs multistep tasks on its own and pulls in a human for steps that need sign-off, while security guardrails applied to individual agents screen for personally identifiable information and attempted prompt injection. The spending controls and the security controls both live in the same management console, North Admin, which is a signal that Cohere sees cost and governance as the two things that block agent rollouts inside large organizations.
Cohere is counting on its work with Nvidia Corp. to lower what North costs to run, saying its models produce more tokens per second per node on Nvidia's Blackwell and Hopper GPUs. Whether that translates into the cost savings enterprises are looking for is likely to depend on how customers mix Cohere's models with ones they bring from outside, since North 2 is model-agnostic. The update also arrives less than three weeks after Cohere signed a merger agreement with German AI developer Aleph Alpha GmbH, a deal that will keep the Cohere name on the combined company.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Ghost AI raised $11 million, led by Andreessen Horowitz with Abstract, Audacious Ventures, Nova and SV Angel…
Moonshot AI closed its final private round at about a $50 billion valuation and is aiming for a Hong Kong IPO…

OpenAI told Fortune it is piloting a 'mission interview' for candidates on certain teams, starting with market…

Reactor said it added Nvidia's venture capital arm and Sapphire Ventures to its cap table, bringing its total…

OpenAI will roll out invisible text watermarks to all ChatGPT and Codex plan users in the EU within weeks, and…

Reflection announced Beam, a 501B-parameter open-weight model with 23B active parameters, claiming parity with…
