> ## Content Index
> Fetch the complete content index at: https://nordisk-ai.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Third Wave of the AI Revolution
- URL: https://nordisk-ai.ghost.io/the-third-wave-of-the-ai-revolution/
- Published: 2025-01-19T21:13:54.000Z
- Updated: 2026-08-27T15:20:46.000Z
- Description: Agentic AI is described as the third wave of the AI revolution and could change how companies handle tasks, from customer service to data analysis. Is this the next big leap for AI - or are there challenges that could slow the development?
- Author: Anders Hagen
- Tags: Application Areas, article, #Migrated-1787841634512, #wp, #wp-post, #Import 2026-08-27 14:41, #Import 2026-09-07 13:41

Agentic AI, or intelligent agents, could be the next big step in the development of artificial intelligence, beyond generative AI. While the technology shares similarities with generative AI, it also offers unique opportunities and challenges.

Marc Benioff, CEO of Salesforce, describes agentic AI as the "third wave" in AI's rapid development. "In just a few years, we've seen three generations of AI. First came predictive models that analyze data, then generative AI with deep learning techniques like ChatGPT. Now we're experiencing a third wave - defined by intelligent agents that can independently handle complex tasks," he recently wrote in the New York Times.

### AI Agents as Digital Coworkers

Intelligent agents function as digital coworkers or customer service representatives who communicate through natural language understanding. These AI systems have the potential to relieve and enhance human capabilities in ways that were previously unthinkable.

Benioff envisions a future where companies can use an "AI workforce" that handles customer interactions, analyzes data, optimizes sales strategies, and performs operational tasks in real time - with minimal human oversight.

### Niche Applications and High ROI

AI agents are designed to handle specific tasks with high efficiency. Their narrow focus gives them advantages that more general AI solutions lack. Sesh Iyer, senior partner at BCG X, argues that "agentic AI will become the next wave of value creation at scale" and sees it as an opportunity to fundamentally reshape processes and generate significant productivity gains.

### Ethics and Governance Critical for AI Agents

Just like generative AI, agentic AI requires clear ethical guidelines and governance structures. This includes:

**Rigorous testing to minimize errors**

**Compliance with rules and fair treatment**

**Establishing governance models to ensure responsible use**

Iyer emphasizes the importance of having the right expertise within organizations to design, build, and manage these systems. The necessary domain knowledge often already exists internally, but training and further development can be crucial to maximizing long-term benefit.

### Differences Between Generative AI and Agentic AI

Agentic AI differs from generative AI in its ability to make decisions autonomously, often without human intervention. David Brault of Mendix explains that generative AI can be used broadly across different industries, while agentic AI is designed for more context-specific environments.

Michael Connell, COO at Enthought, points out that agentic AI is best suited for predictable tasks with low risk or low severity in the event of errors.

## Preparing for the Transition to Agentic AI

To successfully implement agentic AI, companies are advised to start small and scale up gradually. Some steps to consider:

**Identify high-impact use cases, e.g. customer service**

**Conduct pilot projects to test and refine AI agent capabilities**

**Understand the technical platforms that support agentic AI**

Sesh Iyer emphasizes the importance of thinking holistically about the user experience and workflows rather than just focusing on the technology. The goal should be to reduce monotonous tasks, improve productivity, and create better collaboration between humans and machines.

## Data and Oversight Are Critical Factors

Just like with generative AI, data is a critical component for success with agentic AI. Connell emphasizes the importance of:

**Clean, correctly labeled, and relevant datasets**

**Sufficient data volume for training and validating models**

Increased use of agentic AI also requires new oversight models, especially in high-risk environments where traditional supervision may be insufficient. Human oversight therefore remains crucial for managing potential unintended consequences.