Salesforce has introduced two new AI models:xGen-Sales and xLAM—designed to enhance its Agentforce platform, which integrates human and autonomous AI agents to improve business efficiency.
xGen-Sales is a proprietary AI model optimized for sales-related tasks such as generating customer insights, summarizing calls, and managing sales pipelines. It automates routine sales activities, allowing sales agents to focus on more strategic tasks. This model increases Agentforce’s ability to autonomously manage customer interactions, nurture sales pipelines, and support sales teams with greater speed and accuracy.
He xLAM The xLAM (Large Action Model) family features AI models designed to execute complex tasks and trigger actions within enterprise systems. Unlike traditional LLMs (Large Language Models), which are content-based, xLAM models specialize in function invocation, allowing AI agents to act independently, such as initiating workflows or processing data without human intervention. xLAM models range in size and capabilities, from smaller on-device models to larger, more robust models for industrial applications.
Salesforce AI Research developed the xLAM models using APIGena proprietary data generation pipeline that helped improve model performance. Early versions of the xLAM models have already outperformed other large models on key benchmarks. For example, the xLAM-8x22B The model ranked #1 in function calling tasks on the Berkeley leaderboards, outperforming even larger models like GPT-4.
These innovations are designed to enable businesses to scale AI-powered workflows more efficiently. Organizations using these models can now automate complex tasks, improve sales processes, and optimize resource allocation.
Non-commercial xLAM models are available on Hugging Face for community review, while proprietary versions will power Agentforce. xGen-Sales It has completed its pilot phase and will soon be available for general use.