Tame the LLM landscape. We'll clarify key factors (context window size, speed) and present our new CRM benchmark for comparing accuracy, cost, and safety for practical model selection insights.
Important: This session is first come, first seated. You do not have a reserved seat.
Role: Administrator, Architect, Developer
Topic: Artificial Intelligence
Session Summary by Einstein: In a dynamic Salesforce conference session, the lead developer advocate provided practical insights into selecting the right large language model (LLM) for CRM applications, emphasizing the importance of speed, memory, and modality in the decision-making process. The Salesforce CRM benchmark was introduced as a tool for making data-driven decisions when choosing LLMs, with a live demonstration showcasing how to evaluate models based on task complexity, cost, and performance. The session highlighted real-world CRM use cases where LLMs can enhance personalization and problem-solving, such as drafting personalized emails, analyzing sentiment, and summarizing lengthy transcripts. Attendees were guided through a four-step process to identify the ideal LLM for their specific CRM use case, leveraging the benchmark's insights into accuracy, cost, speed, and trust and safety metrics.
