Salesforce’s Benioff Pushes the Agentic Enterprise, But Questions Loom Over Agentforce’s Impact

Marc Benioff is set to detail his bold vision on how he sees AI further transforming organizations—hopefully, he won’t have to scramble to rewrite it. At this year’s Dreamforce, Salesforce’s chief executive will introduce his definition of the “agentic enterprise,” a future in which humans and AI agents form a collaborative ecosystem inside organizations.

“In the agentic enterprise model, AI is a powerful partner. It augments human potential and elevates job satisfaction,” the company explains in a post. “The ultimate goal of an agentic enterprise is to boost efficiency and streamline operations, allowing employees to focus on complex tasks while leaving routine work to intelligent agents.”

Disclosure: I am attending Dreamforce as a guest of Salesforce. The company is covering my hotel and flights. However, Salesforce in no way dictated the content of this post. These are my words.

How we got here shouldn’t be surprising, as Salesforce has been championing this direction for nearly two years, since the launch of its Agentforce platform. It started with a “crawl before you walk” approach, educating customers about the benefits of AI agents and addressing skepticism and concerns around the technology. The company quickly shifted to promoting proactive bots before proclaiming the rise of the digital labor force. And Salesforce isn’t the only one doing it either—every tech provider using AI is following suit.

If you’re looking for a sign that AI innovation is happening rapidly, look no further. Over the course of a matter of months, organizations that elect to be “AI-first” are seeing their workforces disrupted by an influx of agents. In a way, Benioff is giving them a promotion, elevating them from being extensions of human workers to collaborative partners on projects.

Salesforce identifies six characteristics of what makes up an agentic enterprise:

• Autonomous agents: It’s not that organizations use any agent—the bot should be able to act independently, free from any script, and have the (pardon the pun) agency to make decisions and handle complex processes without human intervention
• Goal-driven behavior: Agents should be proactive and have an objective they need to accomplish
• Adaptability and learning: Agents should be capable of adapting to changing environments, understanding past interactions, analyzing new data, and adjusting plans and actions accordingly in real-time to meet objectives
• Orchestrated collaboration: Agents must be able to work with other agents to complete tasks. They must coordinate and communicate with an orchestration tool to resolve issues that may be too complex for a single system to handle.
• Tool integration: Bots must be able to support integration with a company’s technology stack, securely interacting with APIs and databases, as well as third-party developer tools
• Humans must remain in the loop: Regardless of what agents can do for organizations, human workers will be responsible for supervising them. Instead of managing other knowledge workers, humans will become “agent bosses.”

It will be interesting to watch Benioff’s keynote and see how he makes a case that draws customers into his vision. Salesforce has led these organizations to the water; now it has to convince them to drink. But that won’t be easy. Colorful case studies and on-stage testimonies alone can’t fully show how businesses might reap the rewards of this fundamental transformation. And despite Agentforce’s evolution, there are still skeptics about the platform’s ability to deliver real results for users.

However, for companies to undergo the transformation Benioff envisions, they’ll need help from the software they’re using. Salesforce laid the groundwork with Agentforce and embedded the agentic development platform into all its products. And it isn’t stopping there: the company is expanding into new sectors—planning to formally introduce an ITSM offering at Dreamforce, putting it in direct competition with ServiceNow, and having already announced Agentforce Vibes, a vibecoding platform designed to accelerate its ecosystem of AI-powered apps. Together, these moves signal Salesforce’s ambition to make AI agents central not just to customer workflows, but to broader enterprise operations.

We can’t forget that Salesforce also continues to face criticism about Agentforce. Some contend that the platform exacerbates “decision fatigue” because it may not differentiate itself sufficiently from other similar tools on the market today. Others argue the company isn’t providing enough transparency on pricing—a concern one analyst warned me would become a sticking point if Salesforce didn’t address it early. Customers also raised issues about Agentforce’s real-time usage visibility. Lastly, there are posts on Reddit alleging that, for all its grandeur, the platform doesn’t halt hallucinations.

And when it comes to investors, even though Salesforce reports the platform generating $100 million in annualized revenue, at least one Wall Street analyst quips that the amount isn’t “significant enough to move the needle.”

To be fair, Salesforce has addressed some of the feedback it has received. However, it still has more work to do to smooth over Agentforce’s edges if it is to be the core agentic operating system of the enterprise.

Benioff’s charisma and stage presence notwithstanding, can Salesforce commit the necessary energy to convincing companies to embrace this agentic enterprise reality? More than two years into the generative AI era, and based on published studies, business leaders don’t appear to be seeing overwhelming returns on enterprise AI investments. Gartner thinks we’re entering the trough of disillusionment.

Benioff’s agentic enterprise vision is bold, but ambition alone won’t carry the day. Salesforce must convince skeptical customers that Agentforce and its broader AI ecosystem deliver tangible results, not just promises. With competitors circling, mixed signals from early adopters, and enterprise AI still facing a credibility gap, Dreamforce will be another test of whether the company can turn conceptual frameworks into operational realities. Success will require not only charismatic presentations but clear proof that AI agents can truly transform the way organizations work.

Featured Image: Salesforce Chief Executive Marc Benioff discusses the company's Agentforce platform at Dreamforce 2024. Credit: Ken Yeung