Agentic AIs —Are we creating autonomous Employees?

The AI Gaze

The integration of Agentic AI into our workflows and daily routines raises a profound question: Are we creating autonomous employees? As critical as this question is, several opinions have been given on this subject by Top AI professionals, which are quite opinionated. Let me quote some of them below:

• Sam Altman (CEO, OpenAI):
  "Soon, AI agents will be able to do the work of an entire company. The question isn’t whether they’ll replace jobs, but how we’ll redefine work itself."

• Fei-Fei Li (Co-Director, Stanford HAI):
  "AI should be thought of as a tool, not a colleague. It doesn’t have intent or consciousness—it’s a mirror of our own data and decisions."

• Andrew Ng (AI Pioneer, Founder of DeepLearning.AI):
  "Agentic AI is like having a tireless, hyper-efficient worker. But without human oversight, it can also be a bull in a china shop."

• Demis Hassabis (CEO, DeepMind):
  "The next generation of AI won’t just follow instructions—it will reason, plan, and act with a level of autonomy that mimics human problem-solving."

I will be elaborating on this subject in this article—Agentic AI—and leaving you with your own opinions on this popular discussion while trying to explain mine as well.

What are Agentic AIs

Agentic AIs are AI systems that are capable of making autonomous decisions in order to complete a task or reach a goal. These systems execute tasks with minimal human intervention. The most exciting thing about AI agents is that they learn to improve over time.

How Agentic AIs Work

• They are designed to work based on an input/goal from the user, in which the constraints and objectives are determined.

• Large Language models from prebuilt functions process this information and then chain together the tasks required to achieve this goal.

• Then the Agentic AI decides on how to execute these steps or tasks independently, and might require user interaction or clarity if needed.

• The agent will take in data from similar tasks and adjust the workflow as needed by the user.

• As the agent works toward its goal, it applies recorded results and actions into a systemic feedback loop, commonly known as the “data flywheel.” This loop pushes the boundaries of the agent, aiming to improve accuracy and efficiency over time.

Agentic AI Vs AI agents

While the terms ‘AI agents’ and ‘Agentic AI’ can be used interchangeably, there is a significant difference between the two.

AI agents refer to any software/system designed to act based on human instructions. They perform mostly specific or predefined tasks, and have quite a limited ability to adapt as Agentic AIs. A typical example is an AI-powered chatbot.

WHILE

The capabilities of Agentic AIs are quite broader. They can reason, perceive, and solve complex problems or execute difficult tasks even in an unpredictable environment or unexpected conditions. They also offer a higher degree of independence while executing their tasks.

Agentic AI and the concept of autonomous employees

The autonomous abilities of agentic AI evoke the concept of being great employees, but this remains a debatable concern. Many experts envision a future where agentic AI systems become active participants in decision-making because of their abilities to perform complex, multi-step tasks, work tirelessly, integrate into workflows easily, and continuously learn to improve during this process. Some even envision a single individual running an entire company with the support of a network of highly advanced AI agents.

While some describe Agentic AIs as ‘autonomous employees,’ I believe that’s a nuanced and somewhat misleading term. Here’s why…

• Humans still set the goals and strategic objectives for these AI agents.

• The Oversight of skilled humans is still very crucial, especially since AI systems are imperfect—They are still very capable of making mistakes.

• At the moment, AI systems do not possess the rights and responsibilities associated with human employees to perform tasks legally.

• AI Agents work based on patterns, algorithms, and goals for improving performance. They lack genuine understanding, self-awareness, motivation, empathy, or ethical reasoning, which can be found in humans.

Real-world applications

Agentic AI is rapidly moving from theory to real-world deployment, transforming industries by automating complex workflows, making autonomous decisions. Here are some practical applications of Agentic AI across sectors:

• Software Development & IT: Agentic AI can be used as an AI coding assistant for explanation, debugging, and suggesting code. For example, Devin AI is referred to as the ‘First AI software Engineer’ created by Cognitive Labs. It can help in building, debugging, and developing full-stack apps from scratch, and has the capability of doing this autonomously.

• Finance & Trading: Agentic AI can be used as Autonomous trading Agents, like managing stock portfolios and reacting to market shifts. For example, JP Morgan LOXM executes large trades effectively, and we have also seen Darktrace’s Antigena for blocking fraudulent financial transactions.

• Customer Services: This is a very common application; we have seen Autonomous AI agents handling customer complaints, reviews, and queries.

Ethical Concerns.

• Bias: Can they unintentionally reinforce stereotypes?

• Job Displacement: Which industries are most at risk?

• Data Privacy: How much personal data is being acquired?

• Explainability: Can we understand why they make certain decisions?

• Overreliance: Are we becoming too dependent on these tools?

Final Thoughts

I believe we are not creating "employees" in the human sense. Agentic AIs lack consciousness, true understanding, motivation, empathy, ethical reasoning, and the broad adaptability of humans. However, we are building sophisticated systems that will redefine the Job landscape and how work is being done. The focus should be on how we can integrate these systems into our workflows to increase productivity and efficiency.

Thanks for reading.

Coming up next: “Responsible AI: Bias, Privacy & the Battle to Keep AI Safe”

Be sure to subscribe so you don’t miss it.

Don’t forget to like, comment, and share with your friends if you found this helpful.