What's the difference between automation and AI in the design and construction industry?

One of the most common questions we receive is: ‘What’s the difference between AI and automation?’ In short, AI is a form of automation, but not all automated solutions use AI. Think of automation as a house; AI would then be a specific room within that house.

One of the most common methods of automation without AI involves using programming languages to write code, procedures, or functions. In this approach, programmers, by using languages like C or Java, must explicitly define the functionality of each procedure or function.

Let’s say you want to create an app to detect bathroom sinks. Maybe you aim to count the number of bathroom sinks for quantity take-offs and wish to automate the process. In programming, you must explicitly define the criteria determining what a sink looks like. For example,

• “If the outer shape is a rectangle and the inner shape is a circle, mark it as a sink.” 

But what if the internal shape isn’t a circle but rather an oval? You’d have to account for that: 

• “If the outer shape is a rectangle and the inner shape is an oval, mark it as a sink.” 

And what about rectangular sinks? 

• “If the outer shape is a rectangle and the inner shape is also a rectangle, mark it as a sink.” 

And what if the sink is square? You get the idea. As you can see, you need to program for every possible variation explicitly.

How about AI? With AI, particularly the machine learning branch, you can train computer systems without explicitly programming every detail. For instance, you can provide numerous examples of sinks in various forms and shapes and label them as “sink”. Through these examples, the AI algorithm ‘learns’ to recognize what represents a sink.

In simple terms, with traditional programming, you’d create a procedure (like a sink detector). In contrast, with AI, you provide numerous examples of input (sink geometry) and the corresponding output (whether the geometry is a sink or not). The AI then learns the procedure without you having to program it explicitly.

Writing these if-then statements can be limiting, particularly when dealing with numerous nuances in geometry. Considering we have more than 100 shapes and types of sinks, writing and maintaining lines of programming codes for all these variations can be time-consuming. In contrast, with AI, you aren’t constrained in this manner. All you need to do is provide examples. As new examples are introduced, the AI systems adapt, becoming more refined and intelligent over time.

So which one is better? Automation without AI or automation with AI?

It’s important to remember that the goal is to use automation (with or without AI) to address business challenges or create new opportunities for your company. Your choice depends on the return on investment you get out of each method. Sometimes, it might be more straightforward to write explicit procedures. However, when dealing with vast amounts of data, utilizing AI to automate tedious tasks can be beneficial.