Salesforce, a leading enterprise software provider, is reducing its reliance on large language models (LLMs) due to reliability issues. Senior executives, including SVP Sanjna Parulekar, noted a decline in trust in generative AI over the past year. This shift follows layoffs of around 4,000 support staff after the deployment of AI agents. Instead of open-ended generative AI, Salesforce is focusing on predictable, rule-based automation in its Agentforce product, aiming to eliminate the unpredictability associated with LLMs. Technical challenges, such as models misinterpreting complex instructions, have raised concerns, evidenced by customer deployment issues reported by firms like Vivint. CEO Marc Benioff emphasized that Salesforce’s strategic focus is now on data foundations rather than solely on AI. Despite recent stock declines, the company anticipates generating over $500 million in annual revenue from Agentforce, reflecting a critical transition in balancing AI innovation with practical business applications.
Source link
Salesforce Executives Acknowledge Trust Issues Amid AI Strategy Shift Following 4,000 Tech Layoffs
Share
Read more