Salesforce AI Research Launches eVerse: The Simulation Engine Powering Enterprise-Grade Voice Agents

Salesforce AI Research has unveiled eVerse, a groundbreaking simulation environment designed specifically for training enterprise-ready voice and text agents. This platform represents a significant advancement in addressing what researchers call 'jagged intelligence' - the phenomenon where AI systems excel at complex tasks but fail at simple ones, creating unacceptable business risks.

The eVerse platform operates through a continuous loop of synthetic data generation, performance measurement, and behavior optimization. By creating virtual environments that mirror real-world operational complexity - including background noise, diverse accents, spotty connections, and conversational crosstalk - eVerse enables rigorous stress-testing before deployment.

Salesforce has already leveraged eVerse to develop Agentforce Voice, their enterprise voice agent solution. The platform allowed developers to simulate thousands of conversation scenarios, identifying and resolving errors long before reaching production environments. This approach has proven crucial for applications where reliability isn't just preferred but mandatory.

The healthcare sector provides a compelling use case. UCSF Health is currently piloting eVerse to train AI agents for healthcare billing simplification. According to industry data, 60-70% of healthcare contact center inquiries are routine and could be fully handled by AI agents, yet much of this knowledge remains undocumented. eVerse's reinforcement learning from human feedback (RLHF) has achieved up to 88% coverage across both routine and complex tasks in preliminary evaluations.

Madhav Thattai, COO of Salesforce AI, emphasized the platform's transformative potential: 'With eVerse, we were able to explore many nuances of human conversation before Agentforce Voice reached production. This type of rigor is what turns breakthrough research into scalable products and dependable customer experiences.'

The platform represents a critical step toward what Salesforce calls 'Enterprise General Intelligence' (EGI) - AI optimized specifically for business applications that must excel in both capability and consistency. While consumer AI often prioritizes impressive demonstrations, enterprise applications demand predictable performance across every interaction.

Silvio Savarese, Chief AI Scientist at Salesforce, noted: 'Our partnership with UCSF Health demonstrates how applied science translates directly to customer value. When you train agents in environments that mirror real-world complexity, they perform reliably when it matters most.'

As enterprises increasingly depend on AI for critical operations, platforms like eVerse may become essential infrastructure for building the trust required for widespread AI adoption. The ability to simulate, test, and refine AI behavior in controlled environments before real-world deployment could redefine how enterprises approach AI implementation.