Salesforce is developing CRM-specific reasoning models that embed enterprise workflow knowledge directly into AI systems. Created with NVIDIA, Koa is built on Nemotron 3 Super and post-trained using synthetic scenarios modeled on nearly three decades of Salesforce CRM experience. The model is designed to reason through multistep processes, including updating opportunities, routing service cases and scheduling follow-ups, while selecting the appropriate tools for each task. Salesforce says Koa produces three times fewer errors on its CRM benchmark while matching or exceeding leading model performance.
The specialized approach gives companies an alternative to general-purpose models for complex enterprise workflows. Because Salesforce controls Koa’s weights and runs inference within its own infrastructure, organizations can also keep sensitive information within defined trust boundaries. For Salesforce, embedding domain expertise directly into Agentforce can strengthen its AI ecosystem while helping customers automate more complex operational tasks across regulated and commercial industries.
Image Credit: Salesforce/NVIDIA
What Makes This Trend Stand Out
- Domain-specific AI Models
- Enterprise software vendors are embedding industry and workflow expertise into specialized models, creating room for more accurate automation in complex business environments.
- Trusted AI Infrastructure
- Controlled model weights and in-platform inference are reshaping how organizations balance automation with data security, compliance and governance requirements.
- Agentic Workflow Automation
- AI systems that reason through multistep tasks and select the right tools can shift enterprise operations from simple assistance toward autonomous process execution.
Sectors Adopting This
- Customer Relationship Management
- CRM platforms are becoming intelligence layers that automate sales, service and follow-up workflows with deeper context and fewer manual interventions.
- Enterprise Software
- Business application providers can differentiate through embedded reasoning capabilities that understand proprietary workflows instead of relying solely on general-purpose AI.
- Cloud Computing
- Secure cloud infrastructure is enabling companies to run advanced AI inference within defined trust boundaries for sensitive enterprise processes.
