Salesforce CEO Marc Benioff has reignited his feud with Microsoft over AI agents, saying that "Copilot disappointed the world" and that it is "repackaged ChatGPT".
Benioff's comments were delivered to Yahoo Finance at this month's Davos World Economic Forum and continue a war of words between the Salesforce CEO and Microsoft.
Benioff has been critical of Microsoft for years. Salesforce’s CRM product is a direct rival to Microsoft’s Dynamics 365, and Salesforce-owned Slack is a major competitor to Microsoft’s Teams as a collaboration solution.
However, Benioff is particularly scathing of Copilot. In October, he described Microsoft's flagship AI service as being “more like Clippy 2.0” than a “transformational experience”. Benioff took to X to argue that Copilot struggles with accuracy and functionality and “insults” customers by entailing the capacity to build their own LLMs.
Microsoft's launch of Copilot Agents, revealed last autumn but only kicking off with wider access this month, stirred the pot even further, given that Salesforce was an AI agent pioneer in the form of its Agentforce solution. Agentforce's AI bots assume responsibilities and tasks autonomously, such as responding to customer queries and providing product descriptions.
Copilot Agents is pitched as doing something similar across Microsoft's suite of communications, collaboration, and productivity apps, even if it isn't quite a like-for-like competitor in the CX space. Parking agentic AI-shaped tanks on Salesforce's lawn has only exacerbated the feud, however, with Davos 2025 providing a platform to land a few more jabs to Microsoft's ribs.
Benioff said:
Everybody knows that Microsoft has disappointed the world with Copilot, and I mean, that's just a repackaged ChatGPT. You know that."
Microsoft spokesperson Frank Shaw declined to comment on Benioff's remark when approached by Yahoo Finance.
Microsoft has primarily avoided rebuttal to Benioff and Salesforce, preferring silence instead, as illustrated by Shaw's declining to comment. However, there was a recent crack in its stoic armour.
In a recent podcast with venture capitalists Bill Gurley and Brad Gerstner, Microsoft CEO Satya Nadella shared his perspective on the future of business applications in the age of AI in what was widely interpreted as a veiled dig at Salesforce. “The notion that business applications exist—that’s probably where they’ll all collapse in the agent era," Nadella remarked. "If you think about it, they are essentially CRUD databases with a bunch of business logic.”
Agentforce Turns Slack Into An All-In-One Digital HQ, Is Proving A Major Win For Salesforce
Agentforce is proving to be a significant win for Salesforce. In Q3, the company closed 200 Agentforce deals with major new clients such as FedEx, IBM, and Accenture. Overall, it secured 2,000 AI-related deals, signalling a broad adoption of its AI-powered capabilities across industries.
"Well, this is going way faster than anything we've ever done, and it has been amazing, and it's very exciting," Benioff told Yahoo Finance at Davos.
Investor enthusiasm for AI agents has fueled significant momentum for Salesforce, with its stock surging over 25 percent since the debut of its new Agentforce solutions at its Dreamforce event in September.
In September, Slack integrated Agentforce into its platform, elevating the platform well beyond its origin as an enterprise messaging service.
Agentforce, previously known as Einstein Copilot, is an AI-powered solution that enhances productivity by delivering actionable CRM insights and task guidance to its agents. Now integrated into Slack, Agentforce brings generative AI capabilities directly into the collaboration platform. This integration enables users to access real-time insights and task assistance effortlessly, streamlining workflows and boosting overall efficiency.
Meanwhile, its most recent iteration, 'Agentforce 2.0', is already impressing. Rebecca Wetteman, CEO & Principal Analyst at Valoir, told CX Today: "This is all about bringing the pieces together so customers can rapidly get up to speed and experience agentic AI without data science skills, data integration skills, model tuning and testing skills. It's about democratising access to agentic AI and putting it in the hands of people."
