Why Salesforce and Microsoft Are Battling for the Future of AI Agents

You know an emerging technology is getting real when Big Tech firms try to own it. That's what’s happening as Salesforce attempts to seize leadership of the autonomous AI agents market.

AI agents are intelligent systems designed to perform tasks autonomously. At its annual Dreamforce event, Salesforce introduced AgentForce, its autonomous AI agent for the workplace. Afterward, Salesforce CEO Marc Benioff became awfully busy attacking the value of large language models (LLMs) and taking shots at Microsoft for good measure. To wit:

• He recently said LLM-based generative AI assistants, such as Microsoft Copilot, are overhyped. On the Rapid Response podcast, he said, “We may have heard from these AI priests and priestesses of these LLM model companies and Microsoft and others about AI curing cancer and climate change, and we all have to plug into these nuclear power plants to get these data centers. None of this is true.”

• And After Microsoft said it will allow organizations to create their own autonomous AI agents within the Copilot Studio platform, Benioff issued a brutal take-down on X: "Microsoft rebranding Copilot as 'agents'? That's panic mode. Let's be real — Copilot's a flop because Microsoft lacks the data, metadata, and enterprise security models to create real corporate intelligence."

Why would a busy CEO of a multi-billion dollar global Big Tech company lash out at LLMs and Microsoft?

Because Benioff knows how important AI agents are becoming. This means:

• He wants to move the conversation about AI away from LLMs and toward AI agents.

• He sees Microsoft’s foray into AI agents as a threat. If he didn’t, he would have ignored Microsoft’s announcement.

  
  

In doing so, he underlined the importance of AI agents. And propped Microsoft up.

The Importance of AI Agents

The market for autonomous AI agents is experiencing rapid growth. The AI agents market is expected to expand from $5.1 billion in 2024 to $47.1 billion by 2030, reflecting a compound annual growth rate of 44.8%. Salesforce wants a piece of that market.

But the numbers don’t quite capture their potential impact. In fact, AI agents have become a hot topic representing a significant leap in the AI ecosystem. Unlike traditional AI systems that react to commands, AI agents can proactively pursue objectives, learn from their experiences, and even interact with other agents and humans in complex ways.

As a result, AI agents may do mundane, tedious tasks that people don’t want to do – like data entry, managing calendars, automating the creation of reports, and a lot more. In other words, what AI has been promising all along.

This autonomy has sparked both excitement and apprehension.

Venture capitalist Jeremiah Owyang has speculated that AI agents will upend websites because businesses will need to figure out how to re-tool their sites for AI agents doing web searches instead of people. Chris Mattmann, NASA JPL’s former chief technology and innovation officer, echoes the concerns of many by asking if we’re ready for a world in which AI functions without human oversight.

Gartner has identified “Agentic AI” as a top tech trend for 2025, emphasizing its potential to transform enterprise operations. Gartner predicts that by 2028, 33% of enterprise software applications will include agentic AI, enabling 15% of day-to-day work decisions to be made autonomously. However, Gartner warns of the need for governance to prevent risks such as bias and privacy issues. Gartner stresses the importance of establishing guardrails to ensure AI aligns with human values and intentions.

Marc Benioff versus Microsoft

This brings us back to Marc Benioff. He has read the tea leaves. He sees autonomous agents as the future, and he’s placed his bets accordingly. Salesforce has embedded itself in 150,000 organizations. When you’re so deeply embedded across corporate America, your product becomes a Trojan Horse for new technology. For instance, Salesforce AgentForce will deploy autonomous AI agents that enhance customer service, sales, marketing, and commerce operations.

Need to automate routine tasks such as answering customer inquiries, qualifying sales leads, and optimizing marketing campaigns? Hey, there’s an agent for that – AgentForce. Need a more efficient way to get data-driven insights from Salesforce? Well, AgentForce uses the Atlas Reasoning Engine to analyze data, make decisions, and execute tasks autonomously.

Microsoft isn’t taking all of this sitting down. On October 21, Microsoft announced:

• The ability to create autonomous agents with Copilot Studio.
• The introduction of 10 new autonomous agents in Dynamics 365 to build capacity for sales, service, finance and supply chain functions.

Here’s the heady way that Microsoft described on its blog the potential impact of AI agents:

Copilot isyour AI assistant — it works for you — and Copilot Studio enables you to easily create, manage and connect agents to Copilot. Think of agents as the new apps for an AI-powered world. Every organization will have a constellation of agents — ranging from simple prompt-and-response to fully autonomous. They will work on behalf of an individual, team or function to execute and orchestrate businesses process. Copilot is how you’ll interact with these agents, and they’ll do everything from accelerating lead generation and processing sales orders to automating your supply chain.

In effect, Microsoft is going to war with its own Trojan Horse: 60 percent of the Fortune 500 are using Microsoft 365 Copilot (per Microsoft).

Microsoft CEO has avoided a war of words with Benioff. Instead, he’s relying on Jared Spataro to answer Benioff’s dismissive remarks by telling 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 . . . 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.”

Marc Benioff versus LLMs

I don’t know how the war between Salesforce and Microsoft will turn out. However, I believe Benioff is wrong to attack LLM-based AI applications that are based on human prompts. ChatGPT has been in the market for only two years, and already the off-the-shelf generative AI products based on LLMs are changing our lives. In addition, we’re still in the early days of businesses developing their own generative AI models based on private data.

Consider two industries that affect all of us: financial services and healthcare.

• Financial services: LLMs are changing how financial services do their work, with firms like BlackRock using them to analyze market data and identify investment opportunities. BlackRock’s systematic active equity team uses LLMs to process vast datasets, uncovering patterns that traditional methods might miss. Hedge fund Bridgewater Associates, managing a $19 billion portfolio, uses LLMs in its AIA Platform for complex tasks like market research.
  
• Healthcare: hospitals are figuring out how to use LLMs to improve efficiency across the board, from streamlining emergency room intake to improving clinical care. For example, HCA Healthcare in Tennessee is integrating LLMs into its hospitals, focusing on improving time-consuming tasks like documenting notes from conversations with patients during visits.

The above examples illustrate how technology changes the world: one task at a time. Sometimes in visible ways. Oftentimes behind the scenes.

Benioff is correct to pressure tech firms to improve LLM-based AI models. They make mistakes and face criticism over issues like bias and intellectual copyright. But on the other hand, so do AI agents. It’s vital to train them properly and keep humans in the loop to correct errors and prevent bias. But he is overlooking the fact that LLMs learn from their mistakes. They improve. Quickly. If anything, the impact of what LLMs can and will do is understated.