How Does AI Work? What is AI?
How does AI work? It’s about breaking things into pieces that the computer can work with.
So, AI, in general, is a broad term. It’s more of a marketing term than a technical term. This is because there are so many different aspects of the technology field about what AI is.
But when talking about AI, we’re talking about a nonlinear way to explain something. That means when we talk about linear and a standard way, then what we would write is an if statement for example.
What that means is either one or two. When we ask a question, it’s either one way or another way. There are usually only two options or there are six options, but they’re fixed options. But with an AI tool, we’re not looking at a fixed set of options, we’re guessing which option is the most likely. It’s about probability. And that’s the biggest difference between AI and non-AI is. With AI we’re looking for probability. So we’ll say most likely, this is the category we’re looking for.
That’s the reason that we do that is because not all answers are definitive. And when they’re not definitive, that’s where AI helps us.
What AI Provides For Us
So we take that and we break that into all the different components of AI, machine learning is one component of it. And in machine learning, we use models to predict the most likely answer for things. Like for example, it could be a classifier. Is this a bottle or is it a can? That would be a type of classifier. Then, the AI tool would respond with a probability of 85% likelihood it’s a bottle, 15% it’s a can. And so then we know, “Oh, it’s a bottle.” So that’s the difference.
The main thing that AI is providing for us, is this probability or a way of solving problems. They’re using a probability rather than a definitive set of answers. So in the shipping industry, AI works in a lot of different ways.
Sometimes we use natural language processing. Or sometimes we use it for receiving and reading documents. That’s one of the common ones. We’ll try and take a document, for example. It’s like a proof of delivery document or a carrier invoice. And we’ll take that invoice from the carrier and we’ll turn that into something that your TMS can use.
For example, a carrier bill comes in and that comes in via email.
An AI tool can:
- take that email, read that there’s an attachment
- figure out what shipping it belongs to,
- extract the amount from that invoice
- match it up with the amount in your TMS. This is to make sure that the auditing of the invoice matches what the quote was from the carrier
Once those two things match and we have a good shipment, then we can help automate more of the workflow. So that’s one example of how AI works in the shipping industry.
Still Don’t Get It?
One of the biggest drawbacks of AI and how AI works is the fact that it’s such a broad term. And, it’s used in so many different ways. This can be difficult for somebody to understand and to say, “how do incorporate AI in my business?” And if you asked me that, my answer would be, I don’t know.
There are so many different ways to use AI that there’s no direct answer to that. So, that’s one of the challenges is making sure people have an understanding of what AI can do for them. And one of the key ways to do that is to talk with the customer or talk with you and find out what problems are you facing. This will let you know what AI tools to use to solve some of those problems. And sometimes those tools have a very difficult barrier of entry.
For example, IBM’s Watson product. That type of product generally requires some work to get it up and running. So sometimes you’ll need somebody to help out with that. This will bring that product further along so it provides some value. And sometimes you don’t need that type of work. It depends on what the AI product is.