top of page

What is AI?

By Paul T Ayres

Do you know what AI is?

Its uses are growing, but I challenge you to pin down the actual use of AI. The few who have figured it out have kept it pretty close to their chest.


AI is simply data science. It incorporates statistics.  It is data analysis on steroids at the speed of light, which should alone get your attention. Certainly, some quantum computing applications and other large applications might approach our more fantastic interpretations of what AI is. But for most of us, it is simply using already existing automation, then simply collecting more data followed by utilizing good data science applying statistical tools and low-cost data storage.

I was at IBM in Endicott, NY, when the CEO came on the all-company loudspeaker to announce the ‘1-megabyte memory chip.’ Wow, we all thought! Now, what is that today? Maybe one syllable or a second of a song on Spotify?

“The 'data storage' as super inexpensive, the improvement in sensor and data collection options, and massive computing power are the real drivers of this AI tsunami.”

Those of us who studied the logic behind the ‘required’ computer language in our college years, actually, have a bit of an advantage to understand just what AI really is. Or at least, we may.


According to a quick Google search, one of the first definitions that come up is:

A high-level programming language structure that repeats instructions based on the results of a comparison.

  • In a ‘DO WHILE’ loop, the instructions within the loop are performed if the comparison is true.

  • In a ‘DO UNTIL’ loop, the instructions are bypassed if the comparison is true.

At the risk of being overly simplistic, I think this serves to simplify just what AI is.

‘Do Loop’ logic was taught to us freshman engineers in 1980. Although ‘Fortran’ as a programming language is gone, much of the logic that is used to code will not change. Why? Because it's reflective of humans and the way we think.

A huge trick to AI according to my studies I’m currently doing at the Kellogg School of Management – Northwestern University, also watching various online seminars on the subject AI, Generative AI, and machine learning, is 'to formulate questions effectively regarding your data.' Most of the work seems to center on trends that when identified offer some insight into the next steps.

Machine Learning is critical and is based on the questions of the data you formulate, at least initially.


Basically, the conclusions of the data analysis are utilized to adjust the next question to be asked of the data. The advantage is the speed at which this happens (Compare this very high-level definition to what Chat GPT 3.5 says below). The computing power is off the charts. The amount of data that can be utilized blows the mind. 

But where do we get data? Where do we get (and afford) computing power? Do you have these questions? I sure do!


AI, or Artificial Intelligence, refers to the simulation of human intelligence in machines that are programmed to think and learn like humans. It encompasses a broad range of technologies and techniques that enable machines to perform tasks that typically require human intelligence.

These tasks include;

  • problem-solving,

  • speech recognition,

  • learning,

  • planning,

  • perception, and

  • language translation.

There are Two Main Types of AI: Narrow AI and General AI.

  • Narrow AI, also known as Weak AI, is designed to perform a specific task, such as facial recognition or language translation. It operates within a limited context and cannot perform tasks outside its programmed capabilities.

  • General AI, on the other hand, refers to a form of AI with the ability to understand, learn, and apply knowledge across a wide range of tasks, similar to human intelligence. General AI remains a theoretical concept and is yet to be fully realized.

AI applications are diverse and have a significant impact on various industries, including healthcare, finance, education, and manufacturing.

Machine learning, a subset of AI, involves the use of algorithms and statistical models to enable systems to improve their performance on a specific task through learning from data, without explicit programming.

In summary, AI represents the development of computer systems that can perform tasks that typically require human intelligence, and it plays a crucial role in shaping the present and future of technology and innovation.


Machine Learning (ML) is a subset of Artificial Intelligence (AI) that focuses on developing systems and algorithms that enable computers to learn and improve their performance on a specific task without being explicitly programmed. In other words, instead of being programmed with specific instructions, a machine learning system uses data to learn patterns, make predictions, and adapt its behavior over time.

The core idea behind machine learning is to allow computers to learn from experience, just like humans do. The learning process involves exposing the system to large amounts of data, allowing it to identify patterns and relationships within the data. The system then uses these patterns to make predictions or decisions without being explicitly programmed for the particular task.

There are Three Main Types of Machine Learning:

  • Supervised Learning: The algorithm is trained on a labeled dataset, where the input data is paired with corresponding output labels. The algorithm learns to map the input data to the correct output, making predictions on new, unseen data.

  • Unsupervised Learning: The algorithm is given unlabeled data and must find patterns and relationships without explicit guidance. Clustering and association are common tasks in unsupervised learning.

  • Reinforcement Learning: The algorithm learns through trial and error by receiving feedback in the form of rewards or penalties. It aims to maximize cumulative reward over time, adjusting its actions based on the outcomes.

Machine learning is widely used in various applications, including image and speech recognition, natural language processing, recommendation systems, autonomous vehicles, and many more. Its ability to process and analyze large datasets efficiently makes it a powerful tool for solving complex problems and making predictions in diverse domains.


Best to start collecting data. It’s such an elementary comment, but how many of us, especially in small businesses are collecting data now?

I took a ‘4th Generation Industrial Revolution’ course at M.I.T. It was eye-opening. The amount of data that can be collected from a simple spindle rotating on a shaft accumulating fiber thread. How? Sensors. Lots of sensors. Do you have a manufacturing process?

JUST WHAT IS YOUR DATA? Where is it? What do you need to do to collect the right data? What is the right data?

Large companies are down this road a long way. Small companies in mature industries likely are not. The cost burdens of the sensors needed to collect data are the tip of the iceberg. Data storage, third-party data access, and computing power add substantial cost to AI initiatives.

Al and ML are not just for manufacturing processes.

People have measures too. So, service industries have tons of potential for improvement in AI and ML applications as well. You likely record some of the performance matrices in spreadsheets or more automated KPI (key performance indicator) systems. These are used for monitoring workload, schedules, capacity, payroll, and for process improvement at a minimum.

Collect data! In a recent presentation put on by M.I.T. on Al, an executive subject matter guest charged in his company with optimizing and implementing AI made the point it is not about AI directly.

And how that automation plan supports corporate strategy and works to deliver margin. All of us have pressure to make a margin today. Not two years from now. There are tradeoffs to keeping the lights on – that is hitting this month’s margin targets and the long-term viability of the company. Additionally, resources are limited. Most of us are in fact, very small businesses with limited capital.

The Forbes Advisor ( lists some interesting business statistics. Outside of low-cost generative AI like CHAT GPT, most companies are not in a position to capitalize on AI or ML due to its expense and the difficulty we have right now trying to buy the capability.

Just 16% of small businesses have 1 to 19 employees: Most companies are much smaller than 19 people!!!!

In fact, multiple sources validate 99.9% of businesses in the United States are under 250 employees and are defined as 1M to 40M USD in revenue generation as defined by the U.S. Small Business Administration or SBA. If a 40M company has EBITDA (a common financial measure that combines earnings before interest and taxes and also adds depreciation and amortization back to get total EBITDA) at 4M-5M dollars, considered solid, they will have difficulty with a 20M dollar AI or ML implementation project. Massive AI projects simply are not affordable for most American companies right now.

So how do we buy AI?

And when we do, what do we do with our data? Is it safe? Can we have our own data but bring in other data? More on this in future newsletters...

This goes way beyond signing up for open generative AI platforms like CHAT GPT. There are good answers in the affirmative to all of these questions. But the price points are high with real data science tools.

In my M.I.T. seminar the week of December 4th, 2023, one data source says prices are even too high for many Fortune 1000 firms. I can’t validate the subject matter expert guest statement, but I’m not surprised and personally believe him.

New technologies often start off quite pricy, and fall in price over time with added performance. Most sources agree that we’ll see market offerings that are affordable for most businesses.

A few interesting figures from my M.I.T. seminar are below:

  • 61% of the seminar’s participants do not have an AI strategy in place.

  • 36% are experimenting with AI trying to figure out their approach.

  • Only 3% report their AI strategy is mature and executing well.

I don't have the details, but I'll speculate these are large more tech-based companies with the budget and resources to pull this off.

It will determine just what data you’ll have to manipulate with Data Science, AI, and ML. You must concentrate on appropriate data collection, making sure you work within practical budget levels.

Please don’t assume all applications in your business are going to utilize AI. In my studies on this AI subject at the Kellogg School of Management at Northwestern University, the emphasis is also AI and ML is not for every aspect of your business. Stating the obvious, maybe.

Think, what can I automate first. Next, determine if ML can make an impact in terms of real margin increases. It could be automation is only what you need to hit your investment return goals. At Kellogg, we are focusing on what questions we have of the data. Including what questions would our customers have of the data if they had access.


In the seminars I’m absorbing, a consistent theme includes experimenting with the algorithms.

Essentially where you determine the questions of the data and the conditions under which you can use data manipulation to create a margin. This is what it’s about. It gets back to the margin. The classic efficiencies include but are not limited to saving time, saving materials, saving rework, and creating opportunities for additional profitable revenue.

In formulating your questions of the data, you will want to;

  • Establish clear benefits that can be quantified in real dollars hitting your bottom line. If you can’t do this on paper, you likely don’t have a good AI or ML initiative.

  • Use your good business initiative process approval methodologies. 

  • Rigorously scrub your assumptions and create an implementation plan that clearly produces a positive impact on margin as a result of the effort and investment.

In future newsletters, we’ll explore the commoditization pressures of AI on businesses. Especially in a data-sharing environment. Data security is something I humbly recommend you take very seriously. Even with ‘secure third-party services.’

I do recommend organizations;

  1. Jump into their strategic automation plans first by accumulating some background in just what is out there.

  2. Next, clear ties must be established through margin increases.

  3. And finally, the assessment if Al or ML is right for your application must pass the test, again, of more margin creation.

If you’d like to discuss Al and ML applications including how to justify then please contact me on the contact form on my website and we’ll get to work on it. It is an exciting time we live in with huge upside potential.


I wish you the very best on your journey through the new technology jungle we find ourselves in. Look for more AI and ML Newsletters from TheFitProfessional1 Newsletter in 2024.

"Time to get to work!"

26 views0 comments


bottom of page