Artificial Intelligence: Practical Essentials for Management

• Overview of popular AI Techniques

Technique: Classification Techniques

Problem it solves: Predict unknown categorical value based on known data (i.e. which category something belongs to)

Example:

• Would Maya like tea or water today?

• Is a person with annual revenue >$100K more likely to walk or drive to work?

Technique: Regression Techniques

Problem it solves: Predict unknown numeric or continuous values based on known data

Example:

• When is Maya likely to come home today?

• Based on the parameters that have led to past sales deals being won or lost, a software can predict the likely outcome of current sales deals

Technique: Decision Trees, Ensemble Learning (Random Forest & Gradient Boosting)

Problem it solves: Predict unknown categorical or continuous values based on known data, but with greater depth, accuracy and rigour

Technique: Clustering

Problem it solves: Identify the different types of things, people or situations that exist in a given population, & what characteristics differentiate these clusters/groups

Example:

• What are the different types of drinks that Maya will choose from?

• What are the major categories of consumers that a grocery store encounters, based on the purchase volume and timing of its customers

Technique: Association Rule Learning

Problem it solves: Determine what people may prefer if they preferred something else

Example:

• If Maya has tea instead of water, is she more likely to also have a cookie with it?

• If a consumer watches Lord of the Rings, how likely is she to also watch The Hobbit?

Technique: Search Algorithms & Monte Carlo Simulation

Problem it solves: Identify the risks/costs involved with a set of choices & find the optimal one

Example:

• What are the robot's chances of being wrong in its choice of drink to serve?

• Which are the best set of chess moves to play to have the maximum chances of winning in the quickest time?

Technique: Reinforcement Learning

Problem it solves: Solve live, interacting problems to decide which action to take next on-the-go

Example:

• How can the robot walk up to Maya to serve her drink?

• Figure out the best marketing campaign out of 5 different options being shown to the audience on-the-go, by quickly learning which one is resonating the most

Technique: Natural Language Processing

Problem it solves: Read text, analyze it (to check errors, determine personality types of the writer, etc.) and respond

Example:

• How to take voice commands directly from Maya

• Based on the analysis of all emails sent by a customer/employee, determine how friendly she is, or whether her communication style is high or low context, etc.

• What's the core concept behind Large Language Models (LLMs) like ChatGPT

Technique: Deep Learning

• Why is DL growing in importance?

• What is the value it brings?

• What is the underlying concept?

Technique: Deep Learning

Problem it solves: Analyze large amounts of data to try to solve a problem the way human brain does

Example: Is it a picture of a dog, or a wolf that looks like a dog?

Technique: Generative AI, Large Language Models & ChatGPT

Problem it solves: Generate & validate content in real-time to automate and improve day-to-day tasks

Example:

• Craft an email, blog, code etc for a specific purpose or audience type

• Explain a new topic

• Create an image, video, music etc based on certain guidelines

7 Lessons to keep in mind while driving AI Adoption in the organization

• Why do organizations fail in effectively using AI

• Understanding the problem first

• How to determine if AI is needed to solve the problem

• How to decide whether to build or buy an AI solution

• How to identify the right solution

• Importance of data

• Is the data timely, usable, structured, complete, accurate, not biased and enough?

• Data dictionary: How do you achieve data readiness

• Data preprocessing techniques

• Preparing resources to be ready for AI introduction

• Establishing processes to increase AI adoption and effectiveness

• Establishing metrics to measure AI effectiveness & justify the choice of solution

• Different between AI and traditional software

• Factors that influence the financial value of an AI solution or firm

• How to estimate AI valuation

• Best practices for AI strategy formulation

• Why a clear corporate strategy is must

• 7 principles of human-AI work policy

• Establishing processes, governance structures & policy guidelines

• Types of risks with AI

• Ethics & Responsible AI

• Course summary

• Concluding points on AI

• Other courses to take