If Salesforce had hands, it would be rubbing them with excitement right now. Over the past nine months, the company has quietly undergone a radical transformation from a traditional software vendor into a living lab for a new kind of enterprise where humans and AI agents work side by side.
In a wide-ranging conversation with Logan Bartlett from Redpoint Ventures, Salesforce CEO and co-founder Marc Benioff lays out what this shift means. It’s not just a tech experiment. It’s a new model for how modern companies will operate: the Agentic Enterprise.
What is an Agentic Enterprise?
At the core of Benioff’s vision is a simple but powerful idea. AI agents, built on large language models, aren’t replacing humans. They are partnering with them.
These agents are not standalone chatbots. They support customer service, supercharge sales, transform marketing, and even go out into the field with technicians.
“It’s not the end of apps,” Benioff says. “It’s the end of apps acting alone.”
It’s like giving every team an always-on co-worker who remembers everything, never gets tired, and speaks the customer’s language.
AI in Action: Sales and Support, Reinvented
If you still think AI is overhyped, consider this:
- Over the last 26 years, Salesforce accumulated more than 100 million uncalled leads due to lack of manpower.
- Now, their AI-powered sales agents handle over 10,000 leads per week, initiating conversations and converting them into opportunities.
- Their human salespeople work in partnership with these agents.
Meanwhile, in customer support:
- Headcount was reduced from 9,000 to 5,000 without impacting customer satisfaction.
- AI handled 1.5 million customer conversations, and satisfaction scores were equal to those of human agents.
To coordinate it all, Salesforce introduced a new role: the omnichannel supervisor, a kind of digital team lead managing collaboration between humans and agents.
This isn’t science fiction. It’s the next stage in human-computer collaboration.
The Agent is Everywhere
Salesforce is embedding agents across the entire business.
- On the website, a conversational AI helps visitors navigate and take action.
- Inside Slack, dozens of agents handle tasks like renewals, benefits, and HR support.
- In the field, agents assist technicians by surfacing relevant data and helping troubleshoot in real time.
“It’s like Tesla Autopilot,” Benioff notes. “It drives until it hits a situation it can’t handle. Then it hands it off to the human.”
Three Layers of the Next-Gen Stack
Benioff outlines a new architecture for enterprise software:
- Data foundation – powered by Salesforce’s Data Cloud, Tableau, MuleSoft, and soon Informatica.
- Application layer – specialized apps for sales, service, marketing, and more.
- Agentic layer – intelligent agents working on top of the app layer.
All three layers must work together to deliver trusted, secure, and powerful AI.
Measuring the Value of Agents
This isn’t just a cool demo. Salesforce is measuring agent performance with real KPIs.
- In support, 50 percent of conversations are now AI-led, 50 percent human.
- Sales, marketing, and field service are rapidly heading the same direction.
- Workforce structures and KPIs are being redesigned to reflect this new reality.
Managing AI agents is no longer theoretical. It’s becoming part of business operations.
Rethinking Pricing in an Agentic World
With agents replacing or augmenting human work, the traditional per-seat pricing model is evolving.
Salesforce now supports:
- Consumption-based pricing (for data or email volume)
- Conversation-based pricing (number of agent interactions)
- Bundled licenses for full-scale enterprise transformation
“Companies aren’t just buying tools anymore,” Benioff says. “They’re buying transformation.”
The New Generation: Agentic Natives
Benioff also sees a cultural shift underway. Today’s young founders, some just 18 years old, aren’t just digital natives. They are agentic natives. They intuitively understand how to build with AI at the core.
“Not hiring college grads? That’s a huge mistake,” says Benioff. “This is where the insights and energy are coming from.”
Hype vs Reality: Still a Role for Humans
While Benioff is a champion of AI, he pushes back on some of the wilder claims like the “end of apps” or the imminent arrival of AGI.
“LLMs are built on finite algorithms and datasets. Humans are not. We’re infinite. We create immune cells when we’re near sick people. GPT-5 doesn’t do that.”
His point is simple. AI is powerful, but human creativity, empathy, and judgment still matter. The best outcomes come when we work with machines, not against them.
Final Thoughts: Don’t Wait for the Future. Build It.
Benioff’s message is clear. The future is arriving fast, and those who move now will lead. The Agentic Enterprise isn’t a theory. It’s already working at Salesforce, Pfizer, FedEx, and OpenAI.
So if you still think of AI as a chatbot glued to your CRM, you might already be behind.
The winners of tomorrow won’t just use AI. They will redesign their entire business around it.
Human plus Agent equals competitive advantage.
