Salesforce’s Benioff Says Microsoft AI ‘Has Disappointed So Many Customers,’ But Vendor Hits Back

'Microsoft has disappointed so many customers with AI,’ Salesforce co-founder and CEO Marc Benioff said.

Salesforce CEO and co-founder Marc Benioff said earlier this week that Microsoft’s artificial intelligence products and strategy “has disappointed so many customers,” touting his company’s own platform and upcoming “agent” brand of AI products as superior to Microsoft’s Copilots.

Jared Spataro, Microsoft’s corporate vice president for artificial intelligence at work, challenged Benioff’s comments in a statement to CRN.

“Every customer is at a different place in their journey, but overall we are hearing something quite different from our Copilot for Microsoft 365 customers,” Spataro said, a rebuttal of Benioff’s comments Wednesday during the San Francisco-based enterprise software vendor’s latest quarterly earnings call.

“Last quarter alone, we saw a customer increase of over 60 percent, and daily users have more than doubled – a clear indicator of Copilot's value in the market,” Spataro said. “When I talk to CIOs directly and if you look at recent third-party data, organizations are betting on Microsoft for their AI transformation.”

Microsoft AI Vs. Salesforce AI

Microsoft has more than 400,000 partners worldwide, according to the vendor. Salesforce has about 12,000.

During Salesforce’s quarterly earnings call, Benioff said that “so many customers are so disappointed in what they bought from Microsoft Copilots because they're not getting the accuracy and the response that they want” while touting Salesforce’s AI agents approach and Agentforce development platform, expected to become generally available in October.

Salesforce’s agents are autonomous, more accurate and set up in minutes, Benioff said. His goal is to hit 1 billion agents by the end of fiscal year 2026. He forecasted thousands of customers on Agentforce by the beginning of next fiscal year.

Microsoft has done a disservice to the emerging AI field by emphasizing an investment in models over platform, he said. Salesforce’s AI against are “not copilots” Benioff said, describing agents as autonomous, more accurate and are set up in minutes, Benioff said.

Agents also “don't require a conversational prompt to take action,” he said. “ You can do advanced planning, reasoning with minimal human input.”

“It is driving me a little crazy, as you probably heard, like when I meet with these customers and they think, ‘I need to build my own model, I have to train my own model, I have to retrain’ – and they are spending a lot of money on this craziness and it's not working,” Benioff said. “It is a disappointment that Microsoft has convinced so many customers and others to move some of these model companies – who are just regulated at this point to be in commodities – that they have to customize this stuff.”

Benioff continued: “It's not true. It can be done in a platform approach and it's better, easier, lower cost … We'll show it.”

His goal is to hit 1 billion agents by the end of fiscal year 2026. “Just in the trials that we have going on, we have like, 100 million identified or more,” he said. “Call it 200 million.”

He forecasted thousands of customers on Agentforce by the beginning of next fiscal year.

Benioff said that Salesforce customers have “a lot of misconceptions about AI.”

“I have been out there very disappointed with the huge amount of money that so many of these customers have wasted on AI,” Benioff said. “They are trying to DIY (do it yourself) their AI … Customers are spending millions of dollars, but are they really getting the results that they want?”

He compared the trend to the early days of cloud computing and other technologies. “They feel like they have to roll their own, build it themselves, get in the weeds, try to figure out – and they're not going to do it better than we're going to do it,” he said. “We're a professional enterprise software company. That's what we do. And we do it with the trust and scale that they need.”

Salesforce models don’t need training or retraining, he said. “You can use a platform like Salesforce to get the highest efficacy of artificial intelligence, the best capability to fully automate your company, achieve all of your goals, and you can do it with professional enterprise software.”

Both Salesforce and Microsoft have been experimenting with different business models as the AI market emerges.

Salesforce executives said on Wednesday’s call that its Agentforce platform will have a consumption-based model that comes out to about $2 per conversation – different from Einstein’s per-user business model.

Microsoft’s Copilot for Security is consumption based, but products such as Copilot for Microsoft 365 are sold per seat.

During Microsoft’s latest quarterly earnings call in July, CEO Satya Nadella illustrated Copilot’s progress by sharing that Copilot customers with more than 10,000 seats more than doubled quarter over quarter, GitHub Copilot has more than 77,000 organization adopters – up 180 percent year over year – and has an annual revenue run rate of $2 billion.

More than 1,000 paid customers used Copilot for Security and Azure AI customers now number more than 60,000, up nearly 60 percent year over year, Nadella said.