# Salesforce's Koa Model Enhances CRM Efficiency with NVIDIA AI

> Jensen Huang at Dreamforce unveiled Koa, Salesforce’s groundbreaking reasoning model, highlighting AI's potential to revolutionize business operations. He asserted that with AI as a core infrastructure, we can 'know everything and do anything,' while stressing that innovation must prioritize safety.

**Source**: blogs.nvidia.com | **Published**: 2026-09-15 | **Type**: article

## Key Facts

- Salesforce's Koa model, built on NVIDIA, reduces CRM errors by 3x, enhancing operational efficiency.
- Open AI models surged from 30% to 70%, indicating a shift towards customization in enterprise solutions.
- Huang's emphasis on AI safety highlights a competitive edge for companies prioritizing secure innovations.
- The partnership positions NVIDIA as a leader in AI infrastructure, potentially boosting market share in CRM.
- Koa's early adoption by major clients like Formula 1 signals strong demand for advanced AI capabilities.

## Summary

At Salesforce's Dreamforce event, NVIDIA CEO Jensen Huang unveiled a bold vision for the future of artificial intelligence, emphasizing its transformative potential for businesses. This announcement coincided with the introduction of Koa, Salesforce's first reasoning model, developed using NVIDIA's advanced Nemotron 3 Super technology. This partnership marks a significant step in integrating AI into customer relationship management (CRM), highlighting the growing importance of AI as a foundational layer in enterprise operations.

Huang articulated a compelling narrative about the evolution of technology, linking electricity and the internet to the current AI revolution. His assertion that “we can know everything and do anything” encapsulates the ambition driving both NVIDIA and Salesforce. The Koa model, which utilizes a proprietary dataset from nearly three decades of CRM deployments, is designed to enhance decision-making and operational efficiency across various industries.

The emphasis on AI safety was a critical part of Huang's message. He asserted that safety should be the foremost priority in AI development, framing it as an engineering challenge rather than a barrier to innovation. Huang's stance reflects a growing industry consensus that prioritizing safety does not have to come at the expense of rapid technological advancement. His remarks signal a shift in how companies might approach AI deployment, balancing speed with the need for robust safety protocols.

The collaboration between NVIDIA and Salesforce is indicative of a broader trend in which companies are increasingly viewing themselves as AI enterprises. Huang noted that the adoption of open AI models has surged, jumping from 30% to 70% within a year. This shift suggests that businesses are not only leveraging existing models but are also investing in custom AI solutions tailored to their specific needs. As every enterprise becomes more agentic and AI-driven, the competitive landscape will evolve, with companies that effectively harness AI gaining a significant advantage.

Koa's capabilities, which include supervised fine-tuning and reinforcement learning, have already demonstrated impressive performance metrics. In Salesforce's CRM Bench, Koa matches or surpasses leading models in key CRM tasks while achieving three times fewer errors. This performance is crucial as businesses increasingly rely on AI to streamline operations and improve customer interactions. The model is set to enter customer pilots this October, with notable clients such as Formula 1 and UChicago Medicine participating. General availability is anticipated in winter 2026, underscoring the urgency for businesses to adapt to these advancements.

The implications of this development extend beyond Salesforce and NVIDIA. As organizations across various sectors begin to adopt AI-driven solutions, the demand for skilled professionals who can navigate this new landscape will grow. Companies that invest in AI capabilities will likely see enhanced productivity and innovation, while those that lag may struggle to keep pace. The message from Huang and Benioff is clear: businesses must engage with AI technologies or risk being left behind in an increasingly competitive market.

Looking ahead, the integration of AI into enterprise systems will likely redefine operational strategies across industries. Companies that prioritize AI safety and innovation will not only enhance their service offerings but also create new market opportunities. As the landscape shifts, the ability to adapt and innovate will be paramount, positioning AI as a critical driver of future business success. The message is clear: the future belongs to those who embrace AI as a core component of their strategic vision.

## Entities

- **Companies**: NVIDIA, Salesforce, Formula 1, UChicago Medicine, Baxter Credit Union, 1-800Accountant, Engine, Xero
- **Products**: Koa, NVIDIA Nemotron 3 Super, Agentforce
- **Technologies**: artificial intelligence, CRM reasoning model, supervised fine-tuning, reinforcement learning, NVIDIA NeMo RL, NeMo Gym, NeMo AutoModel
- **People**: Jensen Huang, Marc Benioff, Rohan Kumar

## Key Concepts

AI safety, agentic enterprise, open models, customer data privacy, innovation and safety, productivity boost from AI, CRM performance, enterprise AI adoption

## Definitions

- **AI safety**: AI safety refers to the engineering practices and considerations necessary to ensure that AI systems function safely and effectively.
- **agentic enterprise**: An agentic enterprise is a company that integrates AI as a core component of its operations, enhancing its capabilities and decision-making.
- **open models**: Open models are AI models that allow for customization and fine-tuning by users, promoting flexibility and innovation.
- **supervised fine-tuning**: Supervised fine-tuning is a process in machine learning where a pre-trained model is further trained on a specific dataset to improve its performance on particular tasks.
- **reinforcement learning**: Reinforcement learning is a type of machine learning where an agent learns to make decisions by receiving rewards or penalties based on its actions.

## Use Cases

- Powering employee agents in Slack
- Customer pilots in Agentforce
- Improving CRM actions with reduced errors
- Custom AI development by enterprises
- Enhancing customer support workflows
- Synthetic enterprise scenario training

## Frequently Asked Questions

**What is Koa?**

Koa is Salesforce's first CRM reasoning model, designed to enhance customer relationship management through advanced AI capabilities. It was built using NVIDIA's Nemotron 3 Super technology.

**How does NVIDIA ensure AI safety?**

NVIDIA emphasizes that safety is a paramount concern in AI development. They advocate for rigorous engineering practices to ensure that products are functional and safe before release.

**What are the benefits of open models?**

Open models allow companies to customize and fine-tune AI systems according to their specific needs. This flexibility promotes innovation and helps businesses adapt to changing requirements.

**How is Koa being used currently?**

Koa is currently running within Salesforce, powering employee agents in Slack and moving into customer pilots. It is expected to be generally available in U.S. regions by winter 2026.

**What industries are covered by Koa's training corpus?**

Koa's training corpus includes synthetic scenarios across over 14 industries, such as manufacturing, financial services, healthcare, and travel, ensuring its versatility in various applications.

## Links

- [Read on Welcome.AI](https://welcome.ai/content/salesforces-koa-model-enhances-crm-efficiency-with-nvidia-ai)
- [Original source](https://blogs.nvidia.com/blog/jensen-huang-dreamforce/)
- [NVIDIA](https://welcome.ai/company/nvidia): Featured company

---

Source: Welcome.AI | https://welcome.ai/content/salesforces-koa-model-enhances-crm-efficiency-with-nvidia-ai