Salesforce and Nvidia Launch Koa: A New Reasoning AI Model for Enterprise Tasks
At its Dreamforce conference, Salesforce announced the launch of a new reasoning model called Koa in collaboration with Nvidia. Built on Nvidia’s open-weight Nemotron base, Koa is tailored to perform sales, marketing, and customer support tasks more efficiently than traditional frontier models like Claude or ChatGPT. Koa represents an alternative approach to AI development by focusing on enterprise-specific needs rather than general-purpose reasoning. It avoids using actual customer data during training, reducing privacy risks and token costs. The model is integrated into Salesforce’s Agentforce platform as a cost-effective solution for automating complex workflows without relying solely on external providers. Salesforce emphasizes Koa's advantages: it uses fewer tokens to complete tasks, supports internal data governance requirements, and offers an open-weight alternative in contrast with proprietary models from companies like Anthropic or OpenAI. Despite this shift, Salesforce continues partnerships such as its recent ClaudeForce collaboration with Anthropic.
Topics: AI enterprise AI Koa model Nvidia Nemotron Salesforce