• Written by Muhammed Abdulla NC, CEO

A purchasing manager at a large retail organisation checks the inventory levels of a particular product, then queries the purchase order table to determine when the next shipment is due, and finally sends a request to the supplier to expedite the delivery. This series of actions, planned and executed toward a specific goal, is an example of what is known as agentic AI, which refers to a system that plans and takes multi-step action toward a goal.

Definition and Clarification

Agentic AI is a type of artificial intelligence that enables systems to perform tasks that require planning, decision-making, and execution, often in a sequence of steps. This is distinct from other types of AI, such as machine learning, which focuses on pattern recognition and prediction. Agentic AI is concerned with taking action to achieve a specific objective, rather than simply providing insights or recommendations.

The term "agentic" refers to the ability of the system to act independently, making decisions and taking actions based on its own reasoning and goals. This does not mean that the system is autonomous, but rather that it is capable of operating with a degree of autonomy, within the bounds set by its designers and operators.

Distinguishing Agentic AI from Related Concepts

Agentic AI is often confused with other concepts, such as automation and robotics. While these fields do involve the use of machines to perform tasks, they are distinct from agentic AI in that they do not necessarily involve planning and decision-making. Automation, for example, typically involves the use of pre-programmed rules to perform repetitive tasks, whereas agentic AI involves the use of reasoning and problem-solving to achieve a goal.

Another area of confusion is with the concept of "intelligent" systems, which are often touted as being able to learn and adapt on their own. While these systems may be able to learn from data, they are not necessarily agentic, as they may not be capable of planning and taking action toward a specific goal.

Applications in Large Organisations

In a large organisation, agentic AI can be applied to a variety of tasks, such as supply chain management, inventory control, and logistics. For example, a system might be used to plan and optimise the route of delivery trucks, taking into account factors such as traffic, road conditions, and time of day. This can help to reduce costs, improve efficiency, and enhance customer satisfaction.

Agentic AI can also be used to support decision-making in areas such as finance and human resources. For example, a system might be used to analyse financial data and provide recommendations for investment or budgeting, or to help with recruitment and talent management by identifying top candidates and predicting their likelihood of success.

For more information on how agentic AI can be applied in your organisation, you can visit our website at www.asscher.ai to learn more about our products and services.

Limitations and Misconceptions

One common misconception about agentic AI is that it is a replacement for human decision-making and action. While agentic AI can certainly augment and support human capabilities, it is not a substitute for human judgment and oversight. Agentic AI systems are only as good as the data and programming they are based on, and they can make mistakes or encounter unexpected situations that require human intervention.

Another limitation of agentic AI is that it can be difficult to design and implement, particularly in complex and dynamic environments. It requires a deep understanding of the problem domain, as well as the ability to model and reason about the system and its goals.

When Agentic AI is Not the Right Choice

There are certain situations in which agentic AI may not be the best choice, such as when the problem is relatively simple and can be solved through automation or other means. In these cases, the use of agentic AI may add unnecessary complexity and cost.

In other cases, the use of agentic AI may be inappropriate due to ethical or regulatory concerns. For example, in areas such as healthcare or finance, there may be strict rules and guidelines that govern the use of AI and automation, and agentic AI may not be permissible.

If you have questions about whether agentic AI is right for your organisation, you can contact us at www.asscher.ai/contact-us to speak with one of our experts.