SYSTEM NOTICE

Auto translation by AI. Be sure, accuracy, nuances and authorial intent may not be fully reflected.

“Humans are the bottleneck” — Salesforce predicts a future of work where agents and humans collaborate

2026年4月1日 22:43

How will the evolution of generative AI change the way companies operate and the way we work?

Marc Benioff, co-founder and CEO of SaaS giant Salesforce, strongly declared in a recent interview the arrival of the "Agentic Enterprise," where AI agents will enter every aspect of business operations.

In this article, we unravel the next-generation technology trends that investors and business professionals need to know now, from the "Slack-centric AI interface strategy" he discussed to "human-agent collaboration" and the "signs of a resurgence of generalists."

1. All AI will be consolidated into "Slack"

When Salesforce acquired Slack five years ago, few may have understood its true intent. However, that strategy is now extremely clear.

1-1. The "conversational interface" foreseen 10 years ago

According to Benioff, it was the company's Chief Futurist, Peter Schwartz, who strongly pushed for this acquisition. Schwartz is said to have foreseen nearly 10 years ago that "the world is moving toward AI and toward agents. To do that, we need a conversational and open interface."

Today, Slack functions as more than just a chat tool; it acts as a gateway to AI models. Benioff stated, "All of our applications are becoming Slack-first," painting a future where work is completed within the conversational UI of Slack rather than the traditional Salesforce screen (Lightning).

2. Human-agent collaboration and the "bottleneck"

Once countless AI agents are deployed within a company, the way work is conducted will change fundamentally.

2-1. The power of language-wielding agents

Current large language models (LLMs) are adept at processing "language." Therefore, agents are achieving dramatic results in areas such as customer service and coding (as code is now a type of language).

Within Salesforce, 15,000 software engineers are utilizing AI, and Benioff revealed that "the engineering organization has likely improved productivity by more than 30%."

2-2. Humans will be the "bottleneck" and the "supervisor"

However, as AI productivity increases, the "humans" who verify and approve it are becoming the bottleneck. Regarding this point, Benioff points out, "At this stage, that is fine. Because AI still has inaccuracies, it is extremely important for humans to remain in the loop (the cycle of judgment)."

Even in customer service, complex issues that agents cannot handle are handed over to humans, utilizing the "comprehensive judgment" unique to humans. For the time being, we will play the role of scrutinizing AI output and guiding it in the right direction.

3. Changes in organizational strategy: The resurgence of the "generalist"

One of the most exciting changes brought about by the spread of AI is the disappearance of boundaries between job categories.

3-1. The power of AI to cross the walls of expertise

Until now, to be successful in a company, one had to be a "specialist" focused on a specific field. However, with the acquisition of AI as a powerful assistant, the situation has changed completely.

"This is an era where engineering executives can simultaneously be product, design, and marketing executives. The reverse is also true; marketing executives are increasingly becoming capable of handling engineering as well."

Even if you cannot write code, you can use AI to quickly create prototypes. There is no need to wait weeks for a request to be fulfilled by a specialized department, and for "generalists" who can instantly turn their ideas into reality, an environment for creating business has never been more accessible.

3-2. Talent reallocation and fierce competition for acquisition

While automation through AI is progressing, the demand for top-tier AI talent and engineers remains high, and Salesforce is actively recruiting excellent students from top universities like MIT. While some tech companies are conducting layoffs, the reasons vary—ranging from simple cost-cutting to the strain of AI investments—and Benioff warns that "using AI as an excuse is negligence." Organizations are currently in a transitional period of rebalancing their talent to align with the AI era.

4. The importance of data infrastructure and AI "safety"

Finally, what investors must not overlook when evaluating companies is "data" and "safety."

4-1. Without data, AI cannot function

At the foundation of Salesforce's architecture lies a massive LLM, with an "integrated data layer" existing on top of it. As Benioff states, "If data is not organized, AI cannot demonstrate its true power," and the ability to connect a company's unique, high-quality data to AI is the source of competitive advantage.

4-2. AI carries the risk of becoming the "new tobacco"

Furthermore, Benioff has a strong sense of crisis regarding AI safety. Having previously described the dangers of social media as the "new tobacco," he suggests that AI carries similar risks, emphasizing that "safety and growth are two sides of the same coin." Without appropriate guardrails (regulations) from the industry as a whole or from governments, the AI industry cannot achieve healthy growth.

Summary

Marc Benioff's perspective shows that AI is not merely an efficiency tool, but an infrastructure that is rewriting everything from corporate UI and organizational structure to individual career strategies.

A world where humans and AI naturally collaborate on the "Slack" platform, and where anyone can unleash their creativity across the boundaries of specialized fields. The time has come for business professionals to consider how to update their skill sets, and for investors to discern which companies are truly transforming into "agentic enterprises."