Software algorithms that are capable of performing tasks that traditionally require human intelligence, such as visual perception, speech recognition, decision-making and language translation.
AI is an “umbrella” concept that is made up of numerous sub-fields, including machine learning, which focuses on the development of programs that can learn when exposed to new data. When boiled down, a machine learning algorithm learns by self-optimising its parameters to achieve least error outputs to observed training data. This can involve a variety of techniques from linear regression to neural networks
Links about AI
In supply chain, AI would serve in the planning, optimisation and analytical decision making space.
To describe what this tech is capable of think of AI-like-tech as something which would be able to crunch data faster and more comprehensively then ever before. Think of it as an upgrade in what your computer is capable of - it was smart before and now its even smarter because the computer can work out what to do next.
Now imagine how cool this would be: you are planning the despatch plan for the next days transport, there are a thousand possible combinations for each driver, each order, each route. You have a choice, sit there till 2am and painstakingly work out what the best solution is - then come back the next day and do the same. Or, switch on your AI assisted software which works all of this out for you and hands you a report which you can then send out to all of your drivers and managers. You look like a genius, you get home unstressed, life is good.
Fun fact, a good TMS (transport management system) can already do this, so AI stands to do the same job better. Because this new generation of smarts is so smart it can take more data in and produce better results faster. The best part of it is that we are now living in a place where people understand that something can't just be clever it also has to be usable by a non-PHD graduate. So what we get to look forward to is a generation of software which is smarter, more efficient to use, more flexible (online versus having to install it), cheaper and easier to use.
Very exciting, so long as you remember that if someone rocks up to your business selling AI, remember that you need to have good data in order to have a good system. Any system. As much as new capabilities are amazing, you still need to have the basics right.
It's like trying to play professional football while still fumbling the grasp, your not going to make it very far if your still playing with balls while the others are running at full speed and kicking goals.
My point: to make AI work, make your business work first, get serious about cleaning up data, using it and gaining insights now. Then upgrade what your system is capable of.