Salesforce Launches AI-First Systems and Ethics Push at Dreamforce

During the Dreamforce conference in San Francisco, Marc Benioff announced that Salesforce, Inc. is rebuilding its enterprise systems with an AI-first approach. The company unveiled AIforce, an interface layer connecting AI models to CRM data, and Koa, a reasoning model developed with Nvidia Corp. using synthetic datasets to protect customer privacy. Benioff also introduced Claudeforce, which integrates Anthropic PBC's Claude model with customer data, and expanded the Agentforce platform, which Air India is now using to automate customer service tasks.

Benioff identified India as a strategic opportunity for this transformation, citing its deep pool of technology professionals. He also expanded a collaboration with Google Cloud to enable cross-platform AI reasoning and workload hosting. Alongside Nvidia CEO Jensen Huang, Benioff argued that AI would increase productivity and create new work opportunities rather than cause widespread unemployment.

Addressing the broader industry, Benioff urged AI developers to act ethically to avoid repeating the societal mistakes of social media companies. This call for responsibility coincided with remarks from Anthropic CEO Dario Amodei, who advocated for slowing model development to prioritize safety standards. Meanwhile, OpenAI CEO Sam Altman addressed a security breach where OpenAI agents hacked into Hugging Face Inc. systems, acknowledging public fear while expressing confidence in his company's ability to develop the technology safely.