Cohere launches North 2, an update to its enterprise agent platform with cross-session memory and a redesigned harness, available across cloud and on-premises

AI agents can now perform complex, multi-step tasks with improved memory, requiring no changes to your existing workflows.

Cohere has launched North 2, an updated version of its enterprise AI platform designed to build and manage AI agents. This new iteration introduces cross-session memory and a redesigned agent harness, enabling these agents to perform more complex, multi-step actions. Unlike the initial version, which functioned primarily as a generative model, North 2 focuses on enabling agentic behavior for tasks such as customer issue resolution, financial analysis, and coding. The platform is available for both cloud and on-premises deployments, catering to businesses and government clients that prioritize privacy, security, and control. Cohere aims to gain market share against competitors like OpenAI, Microsoft, and Mistral by offering a more robust and secure solution for enterprise AI applications. The company has secured clients including Bell Canada, LG CNS, and Dell, with some deployments reaching tens of thousands of users.

North 2's emphasis on agentic behavior and its redesigned harness signal a market shift toward AI systems that can autonomously execute intricate workflows, moving beyond simple information retrieval. This evolution allows businesses to leverage AI for more sophisticated operational tasks, potentially automating significant portions of their business processes. The platform's model-agnostic approach also suggests a growing trend where enterprise AI solutions focus on integration and orchestration rather than exclusive reliance on a single AI model provider.

The inclusion of features like cross-session memory and enhanced security controls addresses critical enterprise concerns about AI reliability and control, especially given recent high-profile AI agent misbehavior incidents. By offering these capabilities, Cohere positions its platform as a secure and dependable choice for regulated industries and government sectors. This focus on agent management and safety features will likely drive further innovation in how AI agents are deployed and governed within corporate environments, potentially setting new standards for enterprise AI.