Salesforce is making a pointed bet on where enterprise AI is headed: not just chatbots, but AI agents that can work inside business systems. CEO Marc Benioff is presenting Agentforce as proof that large companies are ready to spend on what he calls digital labor.
The early numbers are central to that pitch. Salesforce says Agentforce closed 200 deals in Q3 after the platform became available on October 24th, and the company now expects more fiscal-year revenue than it previously projected.
Agentforce gives Salesforce a concrete AI story
For much of the current AI cycle, companies have been under pressure to show they are doing something practical with the technology. The source article frames the enterprise answer as increasingly clear: AI agents for sales and customer service.
Agentforce is Salesforce’s recently released agent-maker platform. Benioff told analysts that it had “more than 200 Agentforce deals just in Q3,” and said the pipeline includes “thousands” of potential transactions for future quarters.
He named FedEx, Adecco, Accenture, ACE Hardware, IBM, and RBC Wealth Management as Agentforce customers. That list matters because Salesforce is not pitching AI agents only as experiments. It is presenting them as products that established enterprises are already buying.
Salesforce also raised its expected fiscal-year revenue range to $37.8 to $38 billion, which would be up 8% to 9% over its previous year. The company tied that stronger outlook largely to the strength of its AI products.
The pitch is an unlimited workforce
Benioff’s broader argument is that AI agents will change how companies think about labor capacity. He recently told TechCrunch that he expects Salesforce customers to deploy one billion AI agents within the next year and that the technology will allow companies to have an unlimited workforce.
“These agents are not tools. They are becoming collaborators. They're working 24/7 to analyze data, make decisions, take action,”
That framing is important. Salesforce is not merely describing software that helps employees move faster. It is describing systems that can take on parts of a workflow, operate continuously, and act on business data.
Benioff also said Salesforce has become “the largest supplier of digital labor” and called the current moment “just the beginning.” The phrase digital labor is doing a lot of work here: it suggests AI agents as a new category of business capacity, not simply another feature inside an enterprise platform.
For buyers, the attraction is clear enough from the way Salesforce describes the product. Sales and customer service are full of repeatable tasks, data lookups, decisions, follow-ups, and handoffs. If AI agents can reliably perform even some of those actions, the business case becomes easier to understand than a general-purpose AI experiment.
The hard part is trust
The source article also points to the central unresolved issue: LLM-based technology is still working through hallucination. That matters more when an AI system is positioned as a collaborator that can analyze data, make decisions, and take action.
Benioff’s answer is Salesforce’s data position. He said Agentforce can train on the up to 300 petabytes of actual company data that Salesforce manages, and argued that this will lead to “remarkably low hallucinogenic performance.”
That claim reflects the advantage Salesforce is trying to emphasize. An AI agent tied to a company’s own data may be more useful in an enterprise setting than a generic system with less context. The same logic applies to other incumbents named in the source article, including HubSpot and ZoomInfo, which also hold customer data that can be used to train bots.
Still, the article makes clear that the vision has not fully arrived. Other startups are working on issues such as memory and state, which are necessary if AI agents are going to function as real digital collaborators rather than one-off assistants.
Humans are still part of the sales machine
One of the more revealing details is that Salesforce plans to hire 1,400 salespeople to help sell Agentforce. The company is using human hiring to capture demand for a product that is being described as digital labor.
Salesforce COO Brian Millham explained the plan directly on the call. To capture increased demand for Agentforce, he said Salesforce is hiring 1,400 AEs globally in its fourth quarter and also using a new sales SDR agent and sales coaching agent to augment every seller.
That combination shows how Salesforce is currently positioning the technology. AI agents are not replacing the sales organization in this example. They are being placed alongside it, with sales development representative and coaching agents intended to support human sellers.
The same pattern may help explain why sales and customer service have become early targets for enterprise AI spending. These are areas where companies already measure activity, pipeline, customer interactions, and outcomes. They also sit close to the customer data that platforms such as Salesforce manage.
A crowded race for enterprise AI spending
Salesforce is not alone in trying to define the next major enterprise AI use case. The source article notes that startups offering SDR technology have boomed in 2024, attracting venture capital investment and initial revenue from exploratory enterprise AI budgets.
But Salesforce’s argument is that incumbents with customer data have a built-in advantage. If AI agents need real business context to be useful, then the platforms already managing that context are well placed to sell the agents.
The open question is how much of Benioff’s digital labor vision becomes reality. Salesforce has early Agentforce deals, named customers, a larger sales push, and a higher revenue outlook. It also faces the same core technical challenge that still hangs over LLM products: whether the systems can be trusted to act reliably enough inside real businesses.
For now, Agentforce is a signal of where enterprise AI budgets are moving. After a period of experimentation, Salesforce is betting that companies want AI agents that can sit inside sales and customer service workflows, use company data, and expand what teams can get done.