Marc Benioff Fires 4,000 Support Staff at Salesforce as AI Agents Take Over Customer Service

Marc Benioff, CEO of Salesforce, has revealed that about 4,000 customer service jobs will be cut, bringing the company’s support division down from 9,000 to 5,000 employees.  The announcement shows how much Salesforce’s AI-powered Agentforce platform has changed things.

The Rise of AI Agents and Why People Should Be Fired

Agentforce from Salesforce has changed the way customer service is done. It is driven by advanced large language models and agentic AI systems.  These AI agents handle questions, solve problems, and send complicated cases to human agents when needed. Benioff called the increase in speed “the most exciting” thing that had happened in his 26-year tenure.

The layoffs are part of a larger reorganization that will happen by 2025. They come after Salesforce cut over 1,000 jobs in February, which is about 1.5% of its 72,000+ global staff, to focus on AI projects.  By July, Benioff had stopped hiring engineers, customer service reps, and lawyers because AI was making the company more productive.  With this recent cut, the total number of jobs being cut in 2025 has reached about 5,000, which is about 5–7% of the 76,453 people who worked for the company in January 2025.  Human agents handle more complex contacts, while AI handles routine tasks.

Benioff calls the cuts a “rebalancing” that will move money and time to high-value areas like sales of AI and product usage.  Over 800 open roles, many of which are related to AI, are listed internally by Salesforce. The company has hired 2,000 salespeople to push AI products.  But the net decrease in numbers shows a harsh truth: AI is not only adding to jobs, it is actually replacing people who do repetitive, high-volume work.

A Change from Benioff’s Earlier Hope

The layoffs at Salesforce are very different from what Benioff said at the AI for Good Global Summit in July 2025, when he played down worries of mass job losses caused by AI.  At the time, he said that AI would “augment” workers instead of replacing them. He stressed that humans had to be in charge because AI was only about 90% accurate, even with private data.  When leaders like Anthropic’s Dario Amodei said that up to half of entry-level white-collar jobs could go away, he joked, “Maybe they have AI I don’t have.”  Benioff laughed off “scary stories” and said that customers weren’t letting go of employees because of AI.

Just two months later, he admitted that 4,000 jobs had been cut. This shows how quickly agentic AI has grown up, with Salesforce acting as its own “customer zero” for testing.  Benioff says that AI opens up new ways to make money, like calling leads that hadn’t been contacted before. But some critics say that this change of heart makes him less optimistic and shows the human cost of automation.

The Bigger Picture of the Tech Industry

Layoffs reports that over 82,000 tech workers at 192 companies will lose their jobs in 2025 because of AI. Salesforce’s layoffs are part of this bigger wave. As a reminder.  Other companies have done the same thing, like Atlassian (which cut 150 customer service jobs in July) and Klarna (which later hired people again after AI’s “lower quality” work).  Industry studies say that first-contact bot interactions will rise from 15% in 2024 to 85% in 2025, showing a huge rise in the use of AI in customer service around the world.  This trend risks entry-level jobs, especially in support and data entry. Low-skilled workers, like those from poor families who depend on short-term training for $15,000 to $20,000-a-year jobs, are most likely to be affected.

Tech leaders have different points of view.  Jensen Huang of Nvidia says AI will create new jobs that will increase headcount in the long run, but recent data shows that entry-level hiring is down 15% year-over-year.

Workers’ rights and how they affect society

The Worker Adjustment and Retraining Notification (WARN) Act in the U.S. says that companies must give 60 days’ notice before laying off 500 or more workers. Salesforce seems to have followed this rule in previous rounds of layoffs, but there aren’t many specifics about the most recent cuts.  People who lose their jobs at Salesforce face similar problems in places like India, where education programs often fall short in filling the skills gap.

Supporters of workers say that companies like Salesforce should put money into strong upskilling programs to reduce inequality.  Benioff has promised that employees can move up within the company, but past workers say it’s hard for them to reapply for open positions, especially in fields related to AI.  More broad policy changes are becoming popular, like pilots for universal basic income (UBI) or taxes on AI to pay for retraining.  Researchers who used to work for OpenAI think that the UBI could need to hit $10,000 a month for each person to help with widespread displacement.  Lawmakers and unions are also pushing for “human-in-the-loop” rules to make sure AI systems need human control, but rules aren’t keeping up with how quickly AI is being used.

Conclusion

Retraining workers is important but hard, and they need help from both businesses and the government.  As “customer zero” for AI, Salesforce has shown that other companies will follow suit, making it even more important to protect workers ahead of time.

The number of tech jobs lost in 2025 has already surpassed the total number of jobs lost in 2024. This means companies, workers, and policymakers must all work together to get through this change and make sure that the social and economic costs of AI don’t overshadow the benefits, like higher output and better customer outcomes.  For now, the 4,000 Salesforce employees who lost their jobs are a stark warning that the future of work is coming faster than most people thought.