Agentic AI

The workshop began with a comprehensive overview of how rapidly AI models are developing and how they are now seamlessly integrating into virtually every industry.

A major focus of the session was on advanced prompt engineering. Eng. Abdullah detailed how to effectively interact with and guide AI models, explaining various prompting methods, including:

Zero-Shot: Asking the AI to perform a task without providing any prior examples.

Few-Shot: Providing the AI with a few examples to guide its output format and logic.

Chain of Thoughts (CoT): Encouraging the AI to break down complex problems into step-by-step reasoning.

Tree of Thoughts (ToT): Allowing the AI to explore multiple different reasoning paths simultaneously to find the best solution.

The session also covered essential technical aspects of AI deployment. Participants learned how to calculate tokens for different models and manage API usage efficiently. To ground these concepts in reality, Eng. Abdullah shared real-life application examples and case studies from his work with clients in Saudi Arabia.

Additionally, the instructor dedicated time to explaining RAG (Retrieval-Augmented Generation). He clarified that RAG is a powerful technique used to improve the accuracy and reliability of AI models. Instead of relying only on the data the AI was initially trained on, RAG allows the model to search through an external, custom database (like company documents or medical records) to retrieve specific facts before answering a prompt. This ensures the AI's responses are context-specific, up-to-date, and far less likely to generate incorrect information (hallucinations).

One of the most engaging parts of the workshop was the practical demonstration. Eng. Abdullah showcased a real-world workflow using n8n (an automation tool) designed to evaluate training and medical sessions. He demonstrated an automated AI system that can rate session transcripts to determine if a doctor or trainer asked the correct, necessary questions during their interaction with a patient.

The workshop concluded with an open and highly interactive Q&A session. Eng. Abdullah gave participants ample space to ask questions, successfully clarifying many common doubts and misconceptions that people had on their minds regarding the future of AI.

Overall, the session was highly informative, successfully bridging the gap between advanced theoretical AI concepts and real-world business implementation.