Salesforce & NVIDIA Introduce Koa: The First CRM-Specific Reasoning Model for Agentforce

"The most valuable thing Salesforce has built isn't our platform—it's the accumulated knowledge of how enterprise business actually works. With Koa, that knowledge is put inside the model itself. We trained a reasoning engine that understands the structure of a deal, the lifecycle of a service case, and the workflows that vary across industries. That's a different kind of intelligence, and it runs entirely inside your trust boundary."

Why Domain-Specific Reasoning Matters for Enterprise AI

While general-purpose foundation models excel at broad conversational tasks, traditional LLMs frequently struggle with the granular mechanics of complex business systems. Enterprise workflows require deterministic compliance, industry-specific logic, and multi-step execution across connected databases.

Koa addresses this gap by embedding CRM domain expertise directly into the reasoning engine:

• 3x Fewer Errors on CRM Actions: On standard enterprise CRM benchmarks, Koa matches or outperforms leading general-purpose frontier models on core tasks - such as opportunity updates, case routing, and pipeline execution - while producing three times fewer execution errors.
• Synthetic Training Architecture: To maintain absolute data privacy, zero customer data was used in Koa's training corpus. Instead, the model was trained using advanced synthetic scenarios across more than 14 major industries - including financial services, healthcare, manufacturing, and retail.
• Advanced Post-Training: Built in collaboration with NVIDIA, Koa was optimized using Supervised Fine-Tuning (SFT) and reinforcement learning with Group Relative Policy Optimization (GRPO) via NVIDIA NeMo RL, NeMo Gym, and NeMo AutoModel.

Expanding Strategic AI Integrations

Building the Agentic Future with Hajana Technologies