Benioff Dismisses Microsoft Copilot as 'Clippy 2.0,' Touts Salesforce's Slackbot

Marc Benioff, the chief executive of Salesforce, has made a habit of publicly criticizing Microsoft's Copilot, and his latest volley arrived with a pointed historical reference. He called the AI assistant "Clippy 2.0"—a jab at the infamous paperclip-shaped Office helper that became synonymous with unwanted intrusion in the 1990s. The comparison was not flattering. Benioff's message was clear: Microsoft's flagship AI tool, despite its prominence in enterprise software, fails to deliver on its promises.

In a post on social media, Benioff laid out his case with specificity. Copilot, he argued, simply does not work well. The accuracy is poor. Security is compromised—he cited Gartner research indicating that Copilot leaks sensitive data, leaving customers to manage the fallout. To make matters worse, he said, Microsoft then asks those same customers to build their own custom language models to fill the gaps. Benioff claimed he had yet to encounter anyone who experienced a genuinely transformational moment using Copilot or who found value in the costly process of training and retraining custom AI systems. The whole enterprise, in his view, amounted to a modern echo of Clippy's notorious failure to understand what users actually needed.

This is not Benioff's first public critique of Microsoft's AI direction. His comments reflect a pattern of skepticism about Copilot's real-world performance, but they also serve a strategic purpose. Salesforce is making an aggressive move into its own AI ecosystem, and Benioff is positioning the company's alternative as the smarter choice for enterprise customers. The centerpiece of that strategy is a new Slackbot, built on Anthropic's Claude models, designed to operate within Slack—the workplace communication platform Salesforce owns. Benioff describes this tool as the top AI CRM sidekick, framing it as every sales leader's best friend. Unlike Copilot, he suggests, Slackbot actually gets work done. It can tap into customer records, edit data objects, and deploy AI agents to execute sales strategies and generate insights—all without forcing users to abandon their existing workflow.

Benioff's public positioning reflects a broader competitive calculus in enterprise AI. Last year, he abandoned ChatGPT in favor of Google's Gemini 3.0, which he praised effusively for its reasoning capabilities, speed, image quality, and video processing. He described the experience as transformational, claiming that after three years of daily ChatGPT use, two hours with Gemini 3.0 convinced him the technology landscape had fundamentally shifted. That endorsement signaled Salesforce's willingness to shop around for the best underlying AI models rather than lock into any single vendor's ecosystem.

What emerges from Benioff's commentary is a portrait of the current enterprise AI market as fragmented and competitive, with significant performance gaps between offerings. His criticism of Copilot carries weight because it is grounded in specific failures: data leaks, poor accuracy, and the burden placed on customers to fix what the vendor could not. His promotion of Slackbot, meanwhile, emphasizes integration and usability—the idea that AI should work within the tools people already use, not force them to adopt new platforms or spend resources on custom model training. Whether Slackbot lives up to that promise remains to be seen, but Benioff's message to enterprise customers is unmistakable: there are better options than Copilot, and Salesforce has built one.