Cohere has released a new version of its North AI system that lets users deploy rival agents without compromising security, while keeping cost and complexity down.
According to the company, North 2 can now set up AI agents to manage more complex, multi-step processes independently. North 2 also offers reusable agents people can call on, shared knowledge assets an entire organization can draw on, and the ability to retain content from previous sessions to improve automations over time. North also offers a drag-and-drop visual workflow builder, real-time performance monitoring, and integrations with tools already in use.
Security
On security, Cohere says North Admin, a feature that lets users see how their AI is used, also lets teams enforce guardrails at the individual agent level to safeguard data and prevent leaks.
The platform also offers tools to monitor service performance and add detailed logs, autonomy policies designed to take only authorized actions, and ongoing security testing, including red-teaming exercises and third-party vulnerability scanning.
One company currently using North is Bell Cyber, which partnered with the company to bring “government-grade” security to its agentic AI usage.
Interestingly, Cohere also noted that North 2 is “model-agnostic,” so clients can use Cohere’s model or swap to models from other companies to suit their needs. Believe it or not, Cohere actually expects this, as Paul Teyssier, the company’s vice-president of product for AI, told The Logic that the company anticipates clients will keep using its tech for most of their regular AI applications, but could choose alternatives for experimentation and niche tasks (such as if another model outperforms Cohere’s models).
Image Credit: Cohere
News 
News 
Gaming 
News 