Electricity flowing through computer printed circuitboard style brain graphic

Published: November 7th, 2019

 

Artificial Intelligence (AI) projects have become a buzzword catch-all phrase to encompass more a technology trend as opposed to the actual technology. This is not new. I did my PhD over 20 years ago in a field of AI, Expert Systems, when that was the buzzword. We now hardly hear of Expert Systems where human decision processes are encoded into the program logic. Some of its claims were unrealistic just as AI linked concepts such as natural language processing and machine learning are today.

So what is AI? According to Wikipedia, the term “Artificial Intelligence” is often used to describe machines (or computers) that mimic “cognitive” functions that humans associate with the human mind, such as “learning” and “problem-solving”.

There is no doubt it is a growing industry. IDC – a global technology research organization – predicts worldwide annual spending on cognitive and Artificial Intelligence systems will reach $77.6B in 2022. Large organizations spend millions of dollars on AI-based applications but smaller organizations simply can not afford that.

Today’s promises for AI are so great and the AI projects are raising such hopes that in most cases they can not be realistically met. But why can’t they, you ask. The issue is that it is not about just developing the right type of technology. It is also ensuring that the right kind of data has been collected over many years and just as importantly knowing what specific data is needed to support the technology being used to get the desired end results.

So what should you do if you are a small organization and want to use AI? Here are three important considerations.

First, AI’s strength is its predictive capability based on having a great deal of historical data available. You have to know what you want the AI system to predict. For example, future sales of boots depending on weather data.

Second, you have to have sufficient and relevant data to analyze to “feed” the AI system. And you need to have a way of creating baseline values for the prediction based on past data such as previous sales when there was a sudden snowstorm.

Third, you need to have experienced and trained staff to define, create, and maintain the AI program.

Don’t misunderstand me, I’m not saying AI is just hype but I am cautioning particularly for smaller organizations that careful assessment needs to take place before embarking on a project that may not succeed even if the answers to the three considerations above are positive and definitely won’t if the answers are no.

 

 

Share on LinkedIn Share with Google+
More Articles