Is AI replacing engineers? Salesforce says it no longer needs to hire them

Salesforce has said it did not hire any new engineers in fiscal year 2026, with chief executive Marc Benioff attributing the shift to the growing capability of AI-driven coding tools to handle core development work.

Speaking in an interview, Benioff said the company relied on “coding agents” to meet its engineering needs, effectively replacing the need for additional hiring. The comments offer one of the clearest signals yet of how AI is beginning to reshape workforce strategies inside large technology firms.

“I’m not hiring more engineers in fiscal year 26 because I was using coding agents,” Benioff said, adding that AI-driven productivity had provided the “extra capacity” required to deliver work at scale.

---

AI moves from augmentation to substitution

The development reflects a broader shift in how companies deploy artificial intelligence — from assisting employees to potentially replacing certain categories of work. Industry leaders have increasingly suggested that software development, once seen as relatively insulated, may be among the first white-collar functions to see large-scale disruption.

Anthropic chief executive Dario Amodei recently said AI models could automate software engineering “end to end”, reducing the need for human engineers, according to multiple media reports.

At Salesforce, the shift has already extended beyond engineering. The company has also reduced hiring in customer service roles, with AI agents now handling support tasks and lead qualification.

---

Hiring shifts towards revenue roles

While technical hiring slows, Salesforce is expanding in areas where human interaction remains critical. The company has increased its sales workforce by nearly 20%, positioning employees to focus on deal-making and relationship management while AI handles repetitive processes.

The strategy appears aligned with the company’s financial outlook. Salesforce has projected revenue of $46.2 billion and expects more than $16 billion in cash flow, with AI-led productivity gains playing a central role.

Its AI platform, Agentforce, has already grown into an $800 million business, while its broader AI and data segment contributes roughly $2.9 billion in revenue.

Internally, AI systems are being used to manage customer queries, qualify thousands of leads and support deal closures worth millions of dollars.

---

A changing definition of work

Despite the shift, Benioff stopped short of framing AI as a complete replacement for human workers. Instead, he emphasised a hybrid model in which AI augments human capability while changing the nature of roles.

“The future is not about replacing humans completely,” he said, highlighting a growing consensus that AI will reshape, rather than eliminate, most jobs.

However, the pace of change is raising concerns. Protests have emerged in parts of the United States, with activist groups calling for a pause in advanced AI development over fears related to job displacement and long-term risks.

---

Adoption gap remains

Even as companies push AI deeper into operations, adoption remains uneven. A recent Anthropic study found that AI tools can assist with up to 94% of tasks in fields such as coding and mathematics, but actual usage levels are closer to 33%, suggesting a gap between capability and real-world deployment.

Experts say this gap reflects both organisational inertia and the continued need for human judgement in complex or ambiguous tasks.

---

The road ahead

Salesforce’s approach signals a structural shift in workforce planning, where productivity gains from AI directly influence hiring decisions. The company’s model — fewer engineers, more sales roles, and AI embedded across workflows — may become a template for other enterprises navigating similar trade-offs.

As AI systems continue to improve, the question is no longer whether they will change jobs, but how quickly organisations will redesign roles around them. For now, Salesforce’s decision underscores a turning point: engineering capacity is no longer defined solely by headcount, but increasingly by algorithms.