AI Productivity Masterclass

AI Productivity Masterclass. ElEVATE Your PRODUCTIVITY USING AI TOOLs JOHN SOLDATOS, PHD – AI CONSULTANT & TRAINER ([email protected]).

Day 1 (17/03/2025) - Agenda. 01 13:30 - 15:30 Introduction – AI/ML Fundamentals – AI in 2026 02 15:30 – 15:45 Break 03 15:45 – 16:45 Basic Stack for AI Productivity 04 16:45- 17:30 Prompt Engineering.

Day 2 (20/03/2025) - Agenda. 05 13:30 - 15:30 Presentations and Courses with AI 06 15:30 – 15:45 Break 07 15:45 – 16:30 Creative Work with AI Tools & AI Agents 08 16:30- 17:30 AI Use Cases for GEP Group (e.g., Occupational Safety Risk Assessment & Compliance Documentation, Training Content Generation for GEP Academy).

AI Tools to be Presented & Demonstrated. 1 ChatGPT (Writing & Reasoning), Perplexity.aI (Research) 2 Gemini & Google Notebook LM (Content Generation) 3 MS Co-Pilot (Writing, Data Analysis, MS Office 365 Integration) 4 Claude & Claude Skills (AI Skills and Office Automation) 5 Others for Courses & Creative Work (Course.AI, Slides.AI, HeyGen, Synthesia, Automate Video,..).

AGENDA. Introduction. 1. 01. AI/ML Fundamentals – AI in 2026.

01 Introduction.

Introduction to Artificial Intelligence (AI) and Generative AI AI and GenAI Work: Capabilities and Limitations How to use Generative AI Tools to Improve your Productivity.

Understand AI and Ride the wave of AI Achieve considerable Improvement to your Productivity Do more in less time and with better quality Become a better professional.

Findings of MIT Study, March 2023. ChatGPT: The Productivity Advantage.

The BCG & Harvard Study: Similar Findings. A graph showing different skills and a task Description automatically generated with medium confidence.

The Use Cases that Elevate Business Productivity.

The Future of Work. Future of Work | World Economic Forum.

Just Do not die in the next 10 years(!). Powerful AI, if developed responsibly, could radically improve human life Limitations Constraints from Humans: Societal structures, legal requirements (e.g., clinical trials), and human behavior Physical Laws Intrinsic Complexity Speed of the Outside World Need for Data.

Modular Structure – Each Module Focused on Different AI Capabilities and Tools Flexible Learner-Centered Approach Revisit specific parts of the course whenever needed.

AGENDA. Introduction. 1. 01. AI/ML Fundamentals – AI in 2026.

02 AI AND MACHINE LEARNING FUNDAMENTALS.

Exponential growth in computing capacity: Moore’s Law Advances in Semiconductors’ Manufacturing Quantum Computers.

Explosion of data generation: Most of the world's digital data has been produced within just the last few years.

November 2022: The ChatGPT Moment of AI Fastest Growing Application in History Generative AI excels at creating diverse types of content: Images, music, code Future LLMs will only improve.

Instruction-based Computing Users convey programming instructions to the AI Requires Programming Knowledge to make AI systems work Intent-based Computing Users ask what they want using Natural Language Makes AI more accessible than ever before.

Ask Computers What you Want: Intend-Based Computing.

Nowadays there are 100s of Generative AI, LLM-based tools Spanning all different sectors like Marketing, Human Resources, Finance, Healthcare There is virtually an AI tool for everything Learn to find and use the right tool for your Use Case.

Specialized autonomous systems designed to operate within specific domains Many AI Agents are LLM-based Performing tasks with minimal human intervention Characteristics of AI Agents Autonomy Memory Tools Integration Reasoning and Planning.

Comet Browser Eleven Labs Agent Claude Co-Work Agent.

AI Systems Perform Complex Tasks traditionally requiring human-like reasoning, decision-making, or creativity Usually Non-Deterministic and Statistical Try to submit the same prompt to ChatGPT multiple times and compare the outcomes The AI Hype: Non-AI systems are presented as AI systems.

Deep Learning is a subset of Machine Learning that is based on Neural Networks with many hidden layers Benefits of Deep Learning over Machine Learning Significantly Improved performance when large volumes of training data are used No need for feature engineering and selection.

AI vs. Machine Learning vs. Deep Learning. AI and Machine Learning Fundamentals.

Supervised Learning Trains models based on labelled datasets Unsupervised Learning Finding patterns without labelled data and predefined outcomes Reinforcement Learning Trains an AI agent based on rewards and penalties.

Large Language Models (LLMs) are advanced AI systems that are designed to process, understand, and generate human-like text. Key Components of LLMs Transformer Architecture Word Embeddings Training on Large Corpora Billions of Parameters (e.g., GPT-4o has more than a trillion parameters).

LLM: A probabilistic model that assigns a probability P[w1, w2, ..., wn] to every finite sequence w1, ..., wn (grammatical or not) Prediction Model can be very complex e.g., GPT-3 does this with a very large neural network of 175-billion parameters!.

Large Language Models (LLM). AI and Machine Learning Fundamentals.

Do LLMs Understand “Meaning” (?). AI and Machine Learning Fundamentals.

Data used to Train LLMs. AI and Machine Learning Fundamentals.

Next Word Prediction. AI and Machine Learning Fundamentals.

Meanings Change based on context. AI and Machine Learning Fundamentals.

Interacting with LLMs using “Prompt Engineering”.

LLMs and Self-Attention. AI and Machine Learning Fundamentals.

LLMs Fine-Tuning. AI and Machine Learning Fundamentals.

LLMs Factuality. AI and Machine Learning Fundamentals.

Pre-Training Self Supervised Learning Loss Function Optimisation Fine Tuning Reinforcement Learning from Human Feedback.

Generative models (focused on content creation) Example: GPT-4o Reasoning models (focused on logical problem-solving) Example: GPT-o3.

Different Types of AI Models. AI and Machine Learning Fundamentals.

AGENDA. Introduction. 1. 01. AI/ML Fundamentals – AI in 2026.

03 BaSIC AI STACK FOR TASK Productivity and AUTOMATION.

Login Write a Prompt Start with an Action Word Be Specific Add Context Specify Length or Format Iterate and Improve if Necessary.

The quality of responses generated by ChatGPT depends heavily on how good the prompt is Guidelines Clarity and Specificity Tone and Audience Structure and Length Contextual Examples Actionable Insights.

Review and tweak the text Double-check any facts, statistics, or claims made in the blog Consider refining sections to better address your target audience's pain points, interests, or questions Add your unique insights, anecdotes, or branding elements that make the blog stand out and feel authentic Despite this review, ChatGPT has saved you a considerable amount of time.

Prompt ChatGPT to answer a specific email Give Instructions to guide the answer Review and edit the response before sending it to ensure it aligns with your communication style and preferences Give ChatGPT data from your email answers so as to help it answer with your own tone of voice!.

Upload or Paste document (or portions of it) Provide a good prompt for summarization.

Give a Relevant Prompt: “Create a professional and comprehensive customer satisfaction survey….” Give Information about the tone, the types of questions etc..