Artificial intelligence is all around us. When it comes to using this tool in the workplace, it can be hard to know where to start. We sat down with Tony Orzechowski, instructor and retired professional in data analytics, manufacturing, and engineering, to talk about the benefits of AI in the workplace and how we can tailor this technology to fit our needs.
Tony began his career with a bachelor’s degree in mechanical engineering before moving on to Modine Manufacturing, where he worked in processing, manufacturing, engineering, automation, and more. After about five years, Tony’s manager approached him about going back to school in Madison full time to earn his master’s in manufacturing systems engineering. It was in this master’s program that Tony recognized his affinity for analytics, a relatively new focus area for many companies at the time. He was also introduced to artificial intelligence and machine learning. Although it wasn’t called “artificial intelligence” at the time, Tony worked with “expert systems,” pairing manufacturing with algorithms and regression analysis techniques.
Regression Analysis: finding patterns in data and defining their relationships. Regression analytics can be used for predictive purposes in machine learning algorithms – given a certain sentence, regression analysis can predict the next natural word.
When it comes to artificial intelligence and machine learning, Tony has a simple analogy to explain how they relate to one another: “People would say machine learning is artificial intelligence. No, no, no. It’s like saying your engine is a car… Think of [machine learning] as being like the engine in your car… It is extremely important, but it’s not a car, it’s a piece of a car.” Tony goes on to explain, “machine learning can make a prediction, but machine learning doesn’t know what to do with that answer… Now artificial intelligence says, ‘Oh, I know what to do with that,’ and take[s] it further.”
Today, one of the clearest ways people interact with AI is through chatbots. In this context, the machine learning “engine” is called a large language model. Companies such as OpenAI, Google, and others are building these large language model “engines,” which become especially powerful when placed into an artificial intelligence system and interfaced with humans.
Chatbots can be helpful when it comes to workplace productivity, but they need to be given the right context to accurately support you. To help, Tony and other instructors are leading a new training program through UW–Madison that gives working professionals the opportunity to tailor large language models to fit their needs.
Building AI Chatbots Without Code
“The engines are already built. We’re not going to build the engine. What we’re going to do is teach this class how to build everything around the engine… We are providing instructions to this engine, saying ‘Take this question, answer it like this.’”
In this course, participants learn how to define parameters for large language models, identify appropriate data, and build a chatbot suited to their organization’s needs and tone.
“They actually build something that works,” Tony says. “And they’ll be able to deploy it so that somebody else can actually log in and use it.”
Tony has demonstrated this personalization in practice. In one project, he built a chatbot using a specific book as a data source with a cloned version of the author’s voice. The result was a tool that not only provided accurate answers but also reflected the author’s unique perspective. “It’s his voice, it’s his thinking, it has all the right context.”
If you want to confidently join conversations about AI and machine learning, and see what chatbot design can do for your organization, you can learn more and sign up for Building AI Chatbots Without Code.