🏠

Salesforce Koa AI 🚀: Revolutionizing Business Intelligence! 🧠

AI

September 15, 2026 | Author

🎧 Audio Summaries
English flag
French flag
German flag
Japanese flag
Korean flag
Mandarin flag
Spanish flag

🧠Quick Intel


  • Salesforce unveiled Koa, its first reasoning AI model, at Dreamforce.
  • Koa is built upon Nvidia’s open-weight Nemotron model and post-trained for sales, marketing, and customer support tasks.
  • EVP Jayesh Govindarajan stated Salesforce previously relied on external frontier model providers like Claude or ChatGPT.
  • Salesforce partnered with Anthropic to offer ClaudeForcethat allows companies to utilize Claude as their AI interface.
  • Kari Ann Briski, Nvidia’s VP of Generative AI Software for Enterprise, emphasized Koa’s token-efficient inference architecture.
  • Salesforce’s Koa represents the company’s first in-house reasoning model, addressing a prior challenge of lacking a state-of-the-art, American-produced model.

    • đź“ťSummary


    At Salesforce’s Dreamforce conference, a new AI model, Koa, was unveiled. Built upon Nvidia’s Nemotron, Koa represents Salesforce’s first reasoning model. Developed through collaboration with Nvidia, Koa was post-trained for sales, marketing, and customer support applications. EVP Jayesh Govindarajan explained that Salesforce had previously relied on external providers for this capability. Nvidia’s Kari Ann Briski emphasized Koa’s efficient architecture. Simultaneously, Salesforce announced a partnership with Anthropic, offering companies access to Claude via ClaudeForcethat. This development signifies a shift toward American-produced, pre-trained AI models with traceable data origins.

    đź’ˇInsights

    â–Ľ


    KOA: A New Reasoning Model from Salesforce
    Salesforce’s latest innovation, Koa, represents a significant shift in their approach to AI, marking their first foray into reasoning models. Built upon Nvidia’s open-weight Nemotron model, Koa’s development directly addresses the evolving needs of enterprise clients, moving away from the reliance on proprietary, often opaque, AI labs. This strategic move highlights a growing trend where businesses are prioritizing direct interaction with AI models, utilizing their own data and feedback to refine performance – a model that contrasts sharply with the traditional approach of uploading data to external, closed-source systems.

    Integration with Agentforce and Enterprise Use Cases
    Koa’s primary function is to augment Salesforce’s Agentforce platform, providing an alternative reasoning engine for automating tasks like customer service inquiries, appointment scheduling, and sales support. Previously, when an Agentforce agent encountered a complex or multi-step task requiring reasoning, the prompt would be directed to models like Claude or ChatGPT via the Agentforce AI gateway. This new architecture allows for greater control and efficiency within Salesforce’s ecosystem. Jayesh Govindarajan, EVP of Salesforce AI, emphasized this shift, stating that until now, the company had relied on external providers for reasoning capabilities. The team simulated a realistic customer service environment, incorporating both professional and challenging customer interactions, to train Koa effectively. This focus on practical application—simulating irate customers and sales professionals—underscores the model's intended purpose: to excel at the specific tasks Salesforce customers need an agent to perform.

    Technical Foundations and Strategic Partnerships
    The development of Koa leverages Nvidia’s Nemotron model, a key differentiator due to its status as a sovereign, American-trained model with clear data provenance. Kari Ann Briski, Nvidia’s VP of Generative AI Software for Enterprise, highlighted the model’s unique architecture, emphasizing its efficiency in token usage – a crucial factor in managing costs. Salesforce’s strategy also includes a partnership with Anthropic, offering companies access to Claude through ClaudeForce, while maintaining data security within Salesforce’s infrastructure. This dual approach demonstrates Salesforce’s commitment to providing flexible AI solutions, catering to diverse client preferences and technological requirements.

    Related Articles